TAGS ANTI-COLLISION PROTOCOL FOR RFID

0 downloads 0 Views 3MB Size Report
10 Apr 1979 - This technology is expected to improve automation, inventory control, pallet tracking and ... Thus, it is a key issue to develop an efficient anti-collision protocol reducing ..... message, all tags send an answer back to the reader.

Faculty of Engineering Electrical Engineering Department

TAGS ANTI-COLLISION PROTOCOL FOR RFID SYSTEMS

M.Sc. Thesis

Submitted by Mostafa Salah Abd-Elhafez Mohamed Electrical Engineering Department Faculty of Engineering ASSIUT UNIVERSITY Assiut, Egypt 2010

Assiut University Faculty of Engineering Tags

Anti-collision protocol for RFID systems By

Mostafa Salah Abd-Elhafez Mohamed B.Sc., Electrical Engineering Department (Electronics & Communications), Faculty of Engineering, Assiut University, 2003 A THESIS

Submitted in partial fulfillment of the requirement for the degree of

MASTER OF SCIENCE Electrical Engineering Department Faculty of Engineering ASSIUT UNIVERSITY Assiut, Egypt Supervision Committee:

Discussion Committee:

Dr. USAMA SAYED MOHAMMED Prof.Dr. MOHY MOHAMED HADHOUD (Faculty of Engineering, Assiut University)

(Vice President of the University of Menoufiya)

Dr. AMER ABDEL FATAH ALI

Prof.Dr. HANY SELIM GERGES

(Faculty of Engineering, Assiut University)

(Faculty of Engineering, Assiut University) Dr. USAMA SAYED MOHAMED (Faculty of Engineering, Assiut University) Dr. AMER ABDEL FATAH ALI (Faculty of Engineering, Assiut University)

2010

ACKNOWLEDGMENT

To my mother and father

ACKNOWLEDGMENTS In the name of Allah, most Gracious, most Merciful

All deepest thanks are due to ALLAH, the merciful, the compassionate for the uncountable gifts given to me. I

would

like

to

express

my

great

thanks

to

Dr. Usama Sayed in Electrical Engineering Department, Assiut University for his discussions and encouragement. I would like to express my deepest thanks to him for his kind supervision, generous advice, clarifying suggestion and support during each step of this work. Without his help this thesis would not have been what it is. Also, I would like to express my great thanks to Dr. Amer Abdelfatah Ali. I would like to thank all members and friends in the Electrical Engineering Department, Assiut University, for their valuable cooperation that was highly needed during the conduction of this study. I must not forget to express my deepest thanks to my family especially my lovely mother whose prayers, cooperation at all stages of this work and against all odds, have been simply overwhelming.

Mostafa, 2010

ABSTRACT

ABSTRACT

ABSTRACT Radio Frequency Identification (RFID) is a wireless identification system, where the readers try to read the identification (ID) of all existing tags in its range as quickly as possible. The collision problem occurs in signal transmission of the readers or the tags, which leads to slow identification. Limited memory, computational capabilities and lack of internal power source at the passive tag make this problem unique. In this thesis, we tried to address the problem of data collision for tags in RFID systems. The main scope of this thesis is to survey the collision resolution in RFID systems and to propose a new binary tree based protocols that provide better and faster solution. The main idea is to build the protocol with simple dialog between the reader and tags. In this thesis, five protocols are introduced to solve the collision problem. The tags' IDs are the messages to be transmitted which can be used as the splitting rule in the binary tree. The proposed protocols are based on introducing a new identification path based on parallel binary splitting (PBS) technique through the binary tree with minimum prefix and itteration overhead. It is based on dynamically updated tag's relative replying orders. Modifications of the PBS are done to achieve better identification performance. The major advantages of the proposed schemes are low implementation complexity and minimum exchanged bits with minimum overhead and simple logic operation. It consumes the minimum number of transferred bits between the reader and the tags in the identification process. Computer simulations are used to illustrate the performance of the proposed protocols in comparison with the recent schemes in literatures.

I

ABSTRACT

The proposed five protocols can be summarized as follows: i- Parallel Binary Splitting Protocol (PBS): the PBS protocol is based on the parallel binary splitting that dynamically modifies the tag relative reply-order with respect to the adjacent tags. It follows new identification path through the binary tree. It has one-to-one bit simple dialog between the reader and tags. In this protocol, the total number of transmitted bit is equal to twice the binary tree nodes except the leaves nodes. ii- Fast tag anti-collision protocol based on modified PBS technique (FPBS): the FPBS protocol is based on the PBS technique with the assumption of Collision tracking, and the “No Collision” is the default state. The tags send their marked bit in its previously assigned orders until receiving reader acknowledge (notifying) of collision state. The total number of transmitted bits between the reader and tags is equal to the number of the binary tree nodes. The proposed FPBS can provide faster performance in the first and the successive reading rounds. The cost is the assumption of full duplex tag ability. iii- Integrated reader and tag anti-collision protocol based on Similar Topology Trees (STT) in one collision domain: the main idea of this protocol is based on dividing the identification time among the tags and the readers. One bit reply will be sent sequentially in the identification process. The STT technique applies the PBS method to build topology tree configuration for all readers. It provides an integrated

solution

for

collision

environment. II

problems

in

multi-reader

ABSTRACT

iv- Fast and Simple Anti-collision Protocol Based on Up-Down Counter and One Bit Reader Response: the operation of this protocol is based on depth counter and collision pointer embedded inside each tag. The depth counter is used to measure how far the tag position from the current replying tags. However, the collision pointer will keep track the marked bit that will be transmitted when the tag enters the active replying state. The tag will transmit the ID bits without reader interruption until the reader detects the collision state or the identification state. v- Fast anti-collision protocol for moving tags based on Up-Down counter: In this protocol, the modification of the previous protocol is introduced to identify the moving tags. The relative movement between the reader and tags leads to some tags arriving and leaving the reader range. The case is involved in the previous protocol to solve the moving scenario.

III

LIST OF CONTENTS

LIST OF CONTENT

Chapter no.

Title

Page

Chapter 1

INTRODUCTION………………………………...…..................……………

1

Chapter 2

1.1

Background………..….…………………...............……………….

1

1.2

Dissertation layout………………………..……............……..…….

3

RFID SYSTEMS……………………………………….……................….…

5

2.1

Introduction…………………………………...…............………...

5

2.2

RFID History …………………………...…………...........………

6

2.3

RFID System’s Components ……………….……............……….

7

2.3.1

Reader………………………………….............………...

7

2.3.2

Tag……………………………………..........……………

7

2.3.3

Data Processing Subsystem………….........………..........

11

Backscatter (load) Modulation ……….……...……...........……….

12

2.4.1

Tag-Reader coupling principle…………….........……….

12

2.4.2

The mechanism of reader to tag communication................

13

2.5

Data transfer procedures ……………...……............…....................

14

2.6

Manchester Coding………………………….............……………...

15

2.7

RFID frequency bands ……....…………………............………….

15

2.8

Reasons for the increased attention of RFID………............………

16

2.8.1

Motivation for FRID…………………...........………….

16

2.8.2

Advantages of RFID over other auto-identification

2.4

2.9

techniques…………………………….........….…………

17

Examples of recent RFID applications……………...........….…….

18

2.9.1

RFID Tagged Euro Banknotes………………….......……

19

2.9.2

Airbus Signs Contract for High-Memory RFID Tags........

19

2.9.3

Automotive……………………………………….............

19

2.9.4

System for Tracking Health-care Assets………...........….

19

2.9.5

Italian Construction Firm Deploys RFID to Track Offshore Equipment………………………..........……….

20

II

LIST OF CONTENT

Chapter no.

Title

2.9.6

Page

RFID Contains Solution to Chinese Shipping Problems……………………………………..……………

20

2.9.7

RFID Chips in Passports……………………………….…

21

2.9.8

Payment by mobile phones……………………….............

22

Problem of Data Collision in RFID system……………................... 2.10.1 Multi-reader Multi-tag Environment ……………..............

23 23

2.10.2

The Definition of Collision Problem in RFID …………..

23

More challenging problem than multiple-access………………….

24

2.11.1

Constrains on anti-collision protocols in RFID systems....

25

Types of data collisions in RFID system……………………..……

25

2.12.1

Tags collision………………………………………….…

25

2.12.2

Readers collision …………………………………….…..

26

2.13

Performance evaluation of the anti-collision algorithm …………..

28

2.14 2.15

The desirable characteristics of the anti-collision protocol.............. Main system assumptions in building RFID anti-collision protocol

28 29

2.10

2.11

2.12

30

CHAPTER 3

TAG ANTI-COLLISION PROTOCOL……………………..…………….

3.1

Introduction………………………………………………...................

30

3.2

Main classification of tag anti-collision protocols………………..…...

30

3.3

Probabilistic Aloha based protocols…………………………………..

31

3.3.1 3.4

Tree Slotted Aloha (TSA)…………………………………...

32

Deterministic Binary Tree Protocols…………………………………..

34

3.4.1

Query Tree (QT) Protocols……………………………..….

35

3.4.1.1

Basic Query Tree Scheme…………………….....

35

3.4.1.2

Query Tree with Reversed IDs (QTR)………...…

36

3.4.1.3

Query Tree-Based Reservation (RN16QTA)…….

37

III

LIST OF CONTENT

Chapter no.

Title

3.4.1.4

3.5 CHAPTER 4

An Adaptive Memoryless Tag Anti-Collision Protocol for RFID Networks…………………….. 3.4.1.5 Intelligent Query Tree Protocol (IQT)…….……. 3.4.1.6 The collision tracking tree algorithm (CTTA)…... 3.4.1.7 Anti-collision Algorithm based on Binary Tree Searching Algorithms and Balance Incomplete Block Design BIBD (4, 2, 1)..…………………... 3.4.2 Binary Tree Anti-Collision Protocols………………………. 3.4.2.1 A Dynamic Bit Arbitration (DBA) Anti-Collision Algorithm..……………………...……………….. 3.4.2.2 Improvement to the Anti-collision Protocol Specification for 900 MHz Class 0 Radio Frequency Identification Tag……………...…….. 3.4.2.3 ID-Binary Tree Stack Anti-collision Algorithm............................................................... 3.4.2.4 An Efficient Tree-Based Tag Anti-Collision Height-Oriented Protocol…………………...…… 3.4.2.5 Adaptive Binary Splitting (ABS) Tag AntiCollision protocol……………………..…..…….. 3.4.2.6 An Enhanced Anti-collision Algorithm in RFID Based on Counter and Stack (EAA)…………….. 3.4.2.7 Bi-Slotted Tree based Anti-Collision Protocols for Fast Tag Identification………………………. 3.4.2.8 An Anti-Collision protocol of RFID Based on Divide and Conquer Algorithm (DCQT)….…… 3.4.2.9 Novel Anti-Collision Algorithm for Identifying Passive Tags (NEAA)…………………………… 3.4.3 Performance under moving scenario…………….…………. 3.4.3.1 Faster re-identification in successive interrogation sessions…………………………………………... 3.4.3.2 Adaptive binary splitting protocol (ABS)……..… 3.4.3.3 Blocking RFID anti-collision protocol for quick tag identification………………………………… Concluding Remarks………………………………………………….

Page

ANTI-COLLISION PROTOCOLS FOR MULTI-READER ENVIRONMENT……………………………………………………………. 4.1 4.2 4.3

Introduction…………………………………………………..…....... Main classification of reader anti-collision protocols………..…...…. Standard multiple access mechanisms…………………………...……

IV

38 38 40

41 42 43

44 45 46 48 51 54 57 59 61 61 61 62 62 66 66 67 69

LIST OF CONTENT

Chapter no.

Title

4.4 4.5 4.6 4.7 4.8

4.9 4.10 4.11 4.12 4.13 4.14 4.15 4.16

CHAPTER 5

European Telecommunications Standards Institute (ETSI) EN 302 208 (Listen-Before-Talk (LBT))……………………………...….….. Colorwave: An Anti-collision Algorithm for the Reader Collision Problem………………………………………………………....….… Channel Monitoring Algorithm……………………………….…….. Solution to the Reader Collision Problem based on Central Cooperator (CC)-RFID……………..………………………..……... RFID Reader Anti-collision Algorithm Using a Server and Mobile Readers Based on Conflict-Free Multiple Access (CFMA)……………………………………………………….……... Pulse Protocol………………………………………………………. Enhanced Pulse Protocol using Slot Occupied Probability in Dense Reader Environment…………………………………………..……. Pulse Protocol with reduced energy consumption………………........ GENTLE: Reducing Reader Collision in Mobile RFID Networks………………………………………………………..……. DiCA: Distributed Tag Access with Collision-Avoidance Among Mobile RFID Readers…………………………………………..…… Multi-Channel MAC Protocol (MCMAC)……………………...…... A Reader Anti-collision MAC Protocol for Dense Reader RFID System (RAMP)………………………………………………........… An Efficient MAC Protocol for Throughput Enhancement in Dense RFID System (Anti-Collision MAC (ACMAC))……….....................

Page

69 70 71 71

73 74 77 80 82 85 85 86 87

PARALLEL BINARY SPLITTING PROTOCOLS………………………

90

5.1 5.2

90

5.3

Introduction……………………………………………………........... Parallel Binary Tree Splitting Protocol for Tag Anti-collision in RFID Systems (PBS)……………………………………………….............. 5.2.1 Reader operation…………………………………………. 5.2.2 Tag operation…………………………………………….. 5.2.3 Performance analysis…………………………………….. Fast Parallel Binary Splitting Tree Anti-collision Technique (FPBS)...

91 92 93 96 102

5.3.1 5.3.2

Collision tracking assumption…………………………... FPBS operation……………………………………………

103 103

5.3.3

Performance analysis……………………………………..

105

5.3.3.1

The upper limit (bound) of the bit-exchanged between the reader and tags…………………

V

108

LIST OF CONTENT

Chapter no.

Title

5.4

5.3.4 Simulation results………………………………………… 5.3.5 FPBS Performance in successive sessions………………. 5.3.6 Concluding remarks of FPBS protocol………………….. Integrated Reader and Tag Anti-collision Protocol for RFID Systems based on Similar Topology Trees (STT) in the one collision domain………………………………………………………………. 5.4.1 Similar Topology Trees (STT)………………..………… 5.4.2 5.4.3

CHAPTER 6

109 113 115

117 118

Practical assumptions of STT protocol………………… Reader and tag operation based on Similar Tree Topology

118 119

5.4.4 Performance Analysis……………………………………. 5.4.5 Simulation Results……………………………………….. 5.4.6 Concluding remarks of STT reader anti-collision protocol ANTI-COLLISION PROTOCOL BASED ON UP-DOWN COUNTER

121 124 126

AND ONE BIT READER RESPONSE…………………………………….

127

6.1

Introduction……………………………………………………….....

127

6.2

Protocol operation…………………………………………………… 6.2.1 Reader Operation…………………………………………. 6.2.2 Tag Operation…………………………………………….. Performance Analysis……………………………………………..….. 6.3.1 Demonstration example…………………………………. 6.3.2 Another example…………………………………………. 6.3.3 Data overhead per one tag identification………………… Simulation results………………………………………………….…. Performance under moving scenario………………………………..... 6.5.1 Leaving tags……………………………………………… 6.5.2 Arriving Tags…………………………………………….. Chapter conclusion ……………………………..……………………

128 128 128 130 130 133 133 133 138 139 140 141

6.3

6.4 6.5

6.6 CHAPTER 7

Page

CONCLUSION AND FUTURE WORK…...…………..…………….……

142

7.1

Conclusion …………………………………………..………………

142

7.2

Direction for Future Research………………………………..….…...

143

LIST OF PUBLICATIONS………………………………………………………...…..…..

145

LIST OF REFERENCES...……….……………….…………………………………..…...

146

VI

LIST OF FIGURES

LIST OF FIGURES

LIST OF FIGURES Figure no

Title

Page

2.1

Main components of RFID system ……………………………......

7

2.2

Block diagram of RFID passive Tag IC with ٢-KB FeRAM …….

8

2.3

RFID label cross section …………………………………………..

8

2.4

EPC structure ……………………………………………………...

8

2.5

Functional block diagram of RFID passive Tag ……………..……

10

2.6

Master-slave principle between application, reader, and tag ……...

11

2.7

Operating principle of a backscatter transponder………………….. 13

2.8

Amplitude- Modulated Backscattering Signal ………...…...……...

14

2.9

Equivalent circuit of rectifying unit in passive tag...……………....

14

2.10 2.11

Representation of full duplex, half duplex and sequential systems over time………………………………………………………….... 15 Real-time locating system (RTLS) in Chinese shipping system…... 21

2.12

Collision due to tag multiple access problem ……….…………….. 25

2.13

Manchester code…………………………………………………… 26

2.14

Three intersected readers ……………………………………..….... 26

2.15

Mobile RFID reader ………………….……………………………

27

2.16

Types of reader collisions ………………………………..……….

28

3.1

The main tag anti-collision protocol classification.…….…………

30

3.2

The basic slot aloha-based probabilistic algorithm …………..…...…..

31

3.3

Example of TSA protocol execution ……………………….……...

33

3.4

Successful, collided, empty time slots of RN16QTA……………...

37

3.5

IQT protocol achieves saving common prefix bits…………….......

39

3.6

Binary tree representing the EPC number ……………………….... 42

3.7

Representation of an EPC as a tree …………………………..……

VII

44

LIST OF FIGURES

Figure no

Title

page

3.8

ID-binary tree stack anti-collision algorithm………………………

45

3.9 3.10

Bit stream between a reader and tags in the interrogation zone and the flow diagram of searching tags ……………………………… Protocol State diagram…………………………………………….

47 47

3.11

Tag ID pointer …………………………….………………………

47

3.12

ABS anti-collision algorithm……………………………………..

49

3.13

ABS tag operation

50

3.14

Process of EAA identification example…………………………..

52

3.15

Flow Chart of BSCTTA

55

3.16

Search Cost (Bits): BSQTA and BSCTTA ………………………

56

3.17

Identification Performance: BSQTA and BSCTTA ………………

56

3.18

The Possible results of Groups of Collisions for the four Successive Collision bits…………..……………………...……………………..

57

3.19

NEAA tree shows process of NEAA step by step ………………………

60

4.1

Sources of reader collision problem……………………………….. 66

4.2

CC-RFID System Architecture ……………...………….................

72

4.3

An example of a reader network model in CFMA ….…………….

73

4.4

Example of Pulse protocol ……………………………………..….

75

4.5

Example of randomized backoff times in pulse protocol.....………. 76

4.6

PULSE Protocol Flowchart ………..………………..…………….. 76

4.7

Reader collision in the conventional Pulse protocol...……………..

4.8

Mitigating reader collision by Slot Occupied Probability ………… 78

4.9

Identification time of the Enhanced Pulse protocol and conventional algorithms...………………………………………… 80

4.10

System Throughput of the Enhanced Pulse protocol and conventional algorithms …….…………………………………………………… 80

4.11

Example of PULSE operation…………………………………...…

4.12

Example of GENLE protocol operation…………………………… 83

VIII

77

82

LIST OF FIGURES

Figure no

Title

page

4.13

Flowchart of working principle of ACMAC………………………. 89

4.14

Network throughput …………………………….…………………

5.1

The difference between parallel splitting identification path and the depth first search path …………………………………………. 92 PBS procedure executing on the Tag...…….……………………… 95

5.2 5.3 5.4 5.5

89

The diagram of four tags binary tree with the exchanged bit stream ……………………………………………………………... 95 Parallel splitting scan for path exploration………………………… 97

5.7

Total transferred bits vs. the number of the tags (32 bits long for all IDs)……………………………………………………….…….. 99 Identification time of different quantity of tags under 40 kb/s bit rate (ID length =16 bits)…………………………………………… 99 FPBS time diagram……………………………………………….. 104

5.8

The diagram of four tags arranged in binary tree………………….. 105

5.9

The transmitted bit stream between reader and tags for parallel splitting algorithm by PBS & FPBS ………………………………. 106

5.10

The diagram of seven tags arranged in binary tree ………………..

107

5.11

Total transferred bits vs. the number of tags, for random IDs …….

109

5.12

The average number of identified tags per second ………………... 111

5.13

Search cost (Bits): average required bits for one-tag identification. 111

5.14 5.15

Comparing average identifying time of the proposed against recent algorithms………………………………………………………….. 112 The identified tree in the last reading session……………………... 113

5.16

The new tree in the next reading session………………………….

115

5.17

STT Protocol flowchart…………………………………………….

120

5.18

Distribution of three intersected readers with six tags…………….

121

5.19

The complete binary tree of the existing six tags …...…………….. 122

5.20

Three readers are building similar trees ……………………….…..

5.21

Comparison of the network throughput at 50 kb/s bit rate………… 125

5.22

Comparison of the identification time……………………………... 125

5.6

IX

122

LIST OF FIGURES

Figure no

Title

page

6.1

State diagram of the proposed up-down counter protocol ………...

129

6.2

The diagram of four tags binary tree to be identified ………...…… 130

6.3

The order of node exploration……………………………………...

131

6.4

Transmitted bit stream between the reader and tags ………………

132

6.5

The example of seven tags arranged in binary tree to be identified.. 133

6.6

Protocol compare in seven tag identification example…………….

6.7

Total transferred bits vs. the number of tags, for 32 bit ID long…... 135

6.8

Identification Time of different number of tags with 40kb/s bit rate (ID length =16 bits)……………………………………………….. 136 Average identifying time of each algorithm………………………. 136

6.9 6.10 6.11

133

Average identifying time with various numbers of tags (Tag ID length=12bit)……………………………………………………… 137 The binary tree with the leaving tag (B)…………………………... 139

X

LIST OF TABLES

LIST OF TABLES

LIST OF TABLES

Table no.

Title

page

2.1

RFID development history …………………………………..

6

2.2

Comparison of passive and active tags ……………….……..

10

2.3

Read range against frequency range…...……………………..

16

2.4

The desirable characteristics of the anti-collision protocol…..

29

3.1

The tag identification process of Query Tree Protocol ………

36

3.2

The anti-collision process of DBA ……...…………………...

43

3.3

Comparison of querying times of different algorithms………

58

3.4

Comparison among the tag anti-collision protocols………….

63

5.1

The anti-collision process of the proposed PBS algorithm ….

97

5.2

The anti-collision process of the proposed FPBS algorithm....

105

5.3

Estimated number of exchanged bits(for DBA, PBS, FPBS) ..

110

6.1

Exchanged bits compare …………………………………….

133

6.2

Estimated number of exchanged bits(DBA, PBS, Up-Down)..

135

7.1

Comparison of the proposed anti-collision protocols….……..

143

XI

ABBREVIATIONS AND SYMBOLS

ABBREVIATIONS AND SYMBOLS

LIST OF ABBREVIATIONS ABS ACK ACMAC ALOHA ASC Auto-ID BA backoff time BIBD BSCTTA BSQTA CC CDMA CFMA CIMC COR CPR CQ CSMA CTTA CW DBA DCQT DFS DiCA EAA ECB e-passport EPC ETSI FCC FDMA FDX FPBS Gen HDX HF ID IQT ISM

Adaptive Binary Splitting Acknowledgment Anti-Collision MAC computer networking system developed at the University of Hawaii Allocated-slot counter Automatic Identification Blocking ABS protocol The random waiting time delay before entering active state Balance Incomplete Block Design Bi-Slotted Collision Tracking Tree Algorithm Bi-Slotted Query Tree Algorithm Central Cooperator Code Division Multiple access Conflict-Free Multiple Access China International Marine Containers Current Order Register Current Path Register Candidate Queue Carrier Sensing Multiple access collision tracking tree algorithm contention window Dynamic Bit Arbitration Divide and Conquer Query Tree Depth First Search Distributed Tag Access with Collision-Avoidance Enhanced Anti-collision Algorithm European Central Bank Electronic Passport Electronic Product Code European Telecommunications Standards Institute Federal Communications Commission Frequency Division Multiple access Full duplex Fast Parallel Binary Splitting Generation Half duplex High Frequency Identification number (unique) Intelligent Query Tree Protocol Industrial, Scientific, Medical frequency bands XII

ABBREVIATIONS AND SYMBOLS

LAN LBT LF LS MAC MCMAC MP2MP MP2P M-readable NEAA NFC NOC NPC OBCT OCR PBS PSC QTA QTR RAMP RFID RN16 RN١٦QTA RTLS SOP STT TDMA TSA UHF WS € c D P R T Tmin Tw W

Local Area Network Listen-Before-Talk Low Frequency Listening Stage Medium Access Control Multi-Channel MAC Protocol multiple points to multiple points multiple points to one point Multiple-readable Novel Anti-Collision Algorithm Near Field Communication Next Order Counter Next Paths Counter One Bit Collision Timeslot Optical Character Recognition Parallel Binary Splitting Progressed-slot counter Query Tree Algorithm Query Tree with Reversed IDs Reader Anti-collision MAC Protocol Radio Frequency Identification 16-bit random temporary ID Query Tree-Based Reservation Real-Time Location System Slot Occupied Probability Similar Topology Trees Time Division Multiple access Tree Slotted Aloha Ultra High Frequency Waiting Stage LIST OF SYMBOLS Currency symbol Control channel Depth counter collision Pointer Reader Tag Minimum listening time for reader in waiting state for monitoring beacon signal from other reader. Waiting time Worst case XIII

CHAPTER 1 INTRODUCTION

Abstract This chapter serves as an introduction to give the required background and a brief account of the problem that the author is trying to address in this dissertation. The importance and significance of the work are discussed. The areas in which the techniques may be applied are clearly explained. The overview of the dissertation is given at the end of the chapter.

CHAPTER 1

INTRODUCTION

CHAPTER 1 INTRODUCTION 1.1 Background In recent years automatic identification (Auto-ID) procedures have become very popular in many service industries, purchasing and distribution logistics, industry, manufacturing companies and material flow systems. Automatic identification procedures exist to provide information about people, animals, goods and products in transit. [1] In recent years Radio Frequency Identification (RFID) is one technology which developed at a rapid speed that has worldwide applications and it is now considered as a third generation identification technology. RFID is a powerful enabling automatic identification technology with ever widening application and a continuous growth. This technology is expected to improve automation, inventory control, pallet tracking and checkout operations in stores, factories, commerce, logistics, security, access control etc;[2],[3] . Auto-ID (automatic identification) is a core component, for example, of automated inventory control systems and supply chain management. Inventories once taken by hand will be conducted automatically. Continual database updates will better reflect the real world. One day it is conceivable that every man-made object will be labeled with a unique identity associated with a digital entity; [4]. RFID is a contactless (wireless), low cost and low power automatic identification system that uses radio frequency link to extracts identification ID (a unique identification number) from a tag attached to the identified object.

1

CHAPTER 1

INTRODUCTION

RFID overcomes the limitations of other automatic identification applications such as bar codes, magnetic cards and IC cards. RFID has advantages, for example, fast identifying speed, data encryption, longevity, and without the necessity for line of sight. RFID systems consist of networked electromagnetic readers (interrogators) and very small RF chips (tags or transponders), and data management software, where the readers try to identify the tags attached to objects as quickly as possible via wireless communications. The reader is an entity with great computation power and memory, while tags have (very) limited computation resources. However, since the readers or the tags communicate over the same shared wireless channel, the collision problem occurs in signal transmission of the readers or the tags, which leads to slow identification. Collision makes both the communication overhead and the transmission delay of a reader and tags have lost their usefulness. Thus, it is a key issue to develop an efficient anti-collision protocol reducing collisions so as to identify all tags in the interrogation zone. This problem is a special case of the multiple-access communication problem. However, RFID introduce a more challenging aspect to the problem because of the severe cost constraints in implementing these tags in practical applications, each tag is desired to be passive (i.e. no battery requirement) and can only have minimal built-in computing circuitry;[5]. The target of building any anti-collision protocol in all current researches is to improve the tag read efficiency in RFID systems by enabling tags to communicate to the reader as fast and reliably as possible with simple tag logic computaion. In this thesis, the problem of RFID data collision for tags and readers is studied from different view points. The thesis introduces a new, simple 2

CHAPTER 1

INTRODUCTION

and fast solutions based on Parallel Binary Splitting (PBS) technique. As a part of this framework, an integrated solution of the two sources of collision (readers and tags) with minimum overhead information is introduced. The PBS searching path is considered a new identification path through the binary tree that achieves self transmission control and dynamically updates replying orders. It has a parallel tree-scan based on self modified of the tags relative order. The thesis also presents an enhanced version of the counter based anti-collision protocols.

1.2 Dissertation layout The thesis consists of seven chapters: Chapter 1: this chapter introduces the thesis. Chapter 2: RFID System & Collision Problem This chapter provides a background about RFID systems. We shall explain in brief the main components of RFID system, its operation, types of tags, passive tag operation with the load modulation, methods of transferring data and power for passive tags, motivation for passive tag RFID system, and examples of recent RFID applications. Moreover, the description of the problem of data collision in general is introduced in this chapter. It provides the reasons, types (tag and reader collision) and why it represents a more challenging problem than multiple-access problem. Then, it presents the factors that evaluate the performance of the anticollision algorithm and the general assumptions in building anti-collision algorithms.

3

CHAPTER 1

INTRODUCTION

Chapter 3: Tag Anti-collision Protocols This chapter introduces a discussion for the current solutions of existing tag anti-collision protocols for RFID system. It gives the main classification of the followed methods and the previously proposed protocols for solving the problem with showing the main drowbacks that we will override in our protocols to produce simple and faster protocol with minimum overhead. Chapter4:

Anti-collision

Protocols

for

Multi-Reader

Environment It presents a survey for the current reader anti-collision protocols. Chapter 5: Parallel Binary Splitting (PBS) Protocols This chapter introduces in detail the proposed tag anti-collision solution based on the Parallel Binary Splitting (PBS) approach. Then it presents an enhanced Fast Parallel Binary Splitting (FPBS). FPBS Performance in successive reading rounds will be studied. The application of PBS on multi-reader environment presents an integrated reader and tag anticollision protocol based on Similar Topology Trees (STT) in one collision domain. The included computer simulations will be used to show a comparison with the most recent techniques. Chapter 6: Anti-collision Protocol based on Up-Down Counter

and One Bit Reader Response It introduces a simple enhanced counter based protocol. It overcomes the main drawbacks of the previously proposed protocols of counter based. It depends on simple dialog between the reader and tags. Protocol Performance under moving scenario will be studied. Chapter 7: Conclusions and the future works are summarized in this chapter. 4

CHAPTER 2 RFID SYSTEM & COLLISION PROBLEM

Abstract The purpose of this chapter is to provide a background about the basics of RFID system. Some recent RFID applications will be mentioned. The definition of collision problem in RFID system for both tags and readers will be studied in detail. This chapter also presents the factors that evaluate the performance of the anticollision algorithm and the general assumptions and constrains in building anti-collision algorithms.

CHAPTER 2

RFID SYSTEM AND THE COLLISION PROBLEM

CHAPTER 2 RFID SYSTEM AND THE COLLISION PROBLEM 2.1 Introduction Radio frequency identification (RFID) is a new flexible automatic identification technology with ability to wireless communication (read and write data) and without the necessity for direct contact or line-of-sight. [2] It is a low power, low cost automatic identification system that uses radio frequency link to extract identification ID from a tag attached to the identified object. RFID allows traders to identify packages automatically without having to individually inspect each package. Inventories once taken by hand will be conducted automatically. Continual database updates will better reflect the real world. Radio Frequency Identification (RFID) systems are emerging as one of the most pervasive computing technologies due to their low cost and their broad applicability. RFID systems consist of tiny integrated circuits equipped with antennas (RFID tags) that communicate with their reading devices (RFID readers) using radio-frequency waves without line of sight. This creates tremendous opportunities for linking various objects from real world. These objects are numbered, identified, cataloged, and tracked. RFID systems present many advantages and features that cannot be found in other ubiquitous computing environments. RFID communication is fast, convenient and its application can substantially save time, improve services, reduce labor cost, thwart product counterfeiting and theft, increase productivity gains and maintain quality standards. This chapter presents the fundamentals of

RFID systems to provide the basic knowledge and

background for the nature of the system operation. It provides a brief review of its history, importance, advantages over other identification

5

CHAPTER 2

RFID SYSTEM AND THE COLLISION PROBLEM

methods (motivations for RFID), system components, and operation, data and power transfer procedures for passive tags. RFID technology has wide applications in the real world as

public

transportation and ticketing, access control, production control, checkout speed-up in shops, secure operations in dangerous environments, localization of objects (like cars in parking lots, or books in libraries) and people. Some of the more recent applications will be mentioned in Section 2.9 in this chapter.

2.2 RFID History RFID technology was invented in 1948, but it was not commercialized until the 1980s. One of its first known applications was during World War II, when it was used by the British radar system to differentiate between friendly and enemy aircraft with attached radio transponders.

Large

powered RFID tags were placed on friendly aircraft. These tags would give response to identify the carrying aircraft as ‘friendly’ when interrogated by a radar signal. Below, Table 2.1 summarizes the RFID development history. [2] Table 2.1 the RFID development history

Decade 

Event 

1940‐1950 

Radar refined and used, Major World War II RFID invented in 194٥ Early exploration of RFID technology, laboratory explorations Developments of the theory of RFID, start of applications field trials Explosion of RFID development. Testes of RFID accelerate. Commercial applications of RFID Standards developments. RFID widely deployed.

1950‐1960  1960‐1970  1970‐1980  1980‐1990  1990‐2000 

6

CHAPTER 2

RFID SYSTEM AND THE COLLISION PROBLEM

2.3 RFID System’s Components The RFID system mainly consists of three basic components: the tag, the reader and the host data processing system.

Fig. 2.1 main components of RFID system [1]

2.3.1-Reader: consists of a radio frequency module, a control unit, and a coupling element to wirelessly communicate with passive RFID tags at short distances (typically less than 10 m). It provides tags with energy and clock synchronization and read identification information back from tags by detecting the backscatter modulation. It uses anti-collision protocol to regulate the communication session with tags. Readers are fitted with an interface that enables them to communicate with an application subsystem and transfer data between the application software and a tag. 2.3.2. Tag: is made up of a silicon microchip attached to coupling element, such as a coiled antenna that enables them to receive and respond to radiofrequency queries from an RFID transceiver, basic modulation circuitry and non-volatile memory that stores the unique data (ID), as shown in Figure 2.2. Figure 2.3 shows an example of real passive tag label.

7

CHAPTER 2

RFID SYSTEM AND THE COLLISION PROBLEM

Fig. 2.2 Block diagram of RFID passive Tag IC with 2-KB FeRAM.,[8]

Fig. 2.3 RFID label cross section. [9]

EPC structure EPC version

Manufacture ID

Product Type

Item ID

Fig. 2.4 EPC structure

The electronic product code (EPC) identifies the manufacturer, product type, code version and serial number as shown in Figure 2.4. Every tag has a global unique identification code (ID), almost with 96 bit long.

8

CHAPTER 2

RFID SYSTEM AND THE COLLISION PROBLEM

Due to the limitless possibilities and low cost, Radio Frequency Identification (RFID) systems are used in a variety of applications to uniquely identify physical objects. Tags can be classified according to the source of power into: active, semipassive and passive. The tag’s power source determines both its range and cost. [4] Active tags contain an on-board power source, such as a battery, it can initiate their own communications; possibly with other tags. Active tags are continually powered by internal batteries. Consequently, active tags can be read from a greater distance than passive tags. Moreover, active tags have more functions than passive tags, e.g. data storage and sensor capabilities. Tags have to be very simple in order to be as cheap as possible. The lifetime of active tags is over when the battery is exhausted. It is non economical to replace or recharge an active tag battery, because of large number of small and cheap nodes. For these reasons, active tags are not practical for use with disposable consumer products. Semi-passive tag contains its own power source used for its internal control circuitry but not used for transmitter power. Passive tags have no internal power source; it is energized by harvesting energy from the continuous electromagnetic radio waves transmitted by the reader and received by tag antenna. RFID allows wireless powering of the low cost passive tags. It is required to minimize the power consumption of tags, which results in a larger interrogation zone. Passive tags are the cheapest but with the shortest read range of a few meters. It reflects energy from the reader for setting up a communication with the reader.

9

CHAPTER 2

RFID SYSTEM AND THE COLLISION PROBLEM

Fig. 2.5 Functional block diagram of RFID passive Tag [9] Table 2.2 Comparison of passive and active tags [12]

Characteristics  

Passive RFID tag 

Active RFID tag 

Power source  Availability 

Provided by reader Within the field of reader High

built In continuous

Low

High

< 3 meter

> 100 meter

Applicable where tagged items movement is constrained smaller $0.05

Applicable where tagged items movement is variable and unconstrained Larger $10 - $50

Signal strength  (Reader to Tag)  Signal strength  (Tag to Reader)  Communication  range  Applicability in  supply chain  Memory   Expense  

10

Low

CHAPTER 2

RFID SYSTEM AND THE COLLISION PROBLEM

As shown in the more recent tag description in Figure 2.5, the tag acts as a programmable data-carrying device and consists of a coupling element (resonant tuned circuit) and a low-power CMOS IC. The IC chip contains an analog RF interface, antenna tuning capacitor, RF-to-dc rectifier system, digital control and electrically erasable and programmable read-only memory (EEPROM), and data modulation circuits. [9] RFID involves contactless reading and writing of data into an RFID tag's nonvolatile memory through an RF signal. Passive tags rely only on RF energy induced by the electromagnetic waves emitted by the reader. In a typical communication sequence, the reader emits a continuous radio frequency wave. When a tag enters in the RF field of the reader, it receives energy from the field. It is powered on by induction and sends the stored ID code or other data from its inner memory to the reader. Table 2.2 shows a comparison of passive and active tags. [12] The data can be exchanged when the passive tag comes in proximity to the reader signal. The read range depends on the trade-off between number of engineering factors: the frequency of RFID system operation, the power of the reader, and interference from other RF devices. To meet cost, size and lifetime requirements, passive tags seem to be the best solution for RFID systems. 2.3.3 Data Processing Subsystem, which can be an application or database, depending on the application. Figure 2.6 presents the master-slave model.

Fig. 2.6 Master-slave principle between application, reader, and tag. [7]

11

CHAPTER 2

RFID SYSTEM AND THE COLLISION PROBLEM

Reader can identify all tags within the interrogation zone of its antenna via wireless RF communication technology and algorithms. Readers can be connected to a host through interfaces and the stored information from the tags can be accessed with data processing systems. Additionally, when linked with a LAN, the processed data can be transmitted over the internet and shared with other network applications. [16] While the reader plays as master of the tags, the reader plays as slave with respect to the application.

2.4 Backscatter (load) Modulation: Communication by means of reflected Power can be regarded as a milestone of the RFID technology. [35] 2.4.1 Tag-Reader coupling principle The tag receives both information and operating energy from reader RF signal. The tag responses by modulating the reflection coefficient of its antenna, by repeatedly shunting the tag coil through a transistor, the tag can cause slight fluctuations in the reader’s RF carrier amplitude. The RF link behaves essentially as a transformer; as the secondary winding (tag coil) is momentarily shunted, the primary winding (reader coil) experiences a momentary voltage drop, thereby backscattering the continues-wave (CW) RF signal transmitted by the reader, instituting half-duplex communication . The reader must peak-detect this data at about 60 dB down. To offer an analogy for the passive powering process, one may think of readers as “shouting” out to passive tags, then extracting data from the resultant echoes. Passive tags are completely inactive in the absence of a reader. [1, 4]

12

CHAPTER 2

RFID SYSTEM AND THE COLLISION PROBLEM

2.4.2 The mechanism of reader to tag communication • The reader continuously generates an RF carrier sine wave, watching always for modulation to occur. Detected modulation of the field would indicate the presence of a tag. • A tag enters the RF field generated by the reader. Once the tag has received sufficient energy to operate correctly, it divides down the carrier and begins clocking its data to an output transistor, which is normally connected across the coil inputs. • The tag’s output transistor shunts the coil, sequentially corresponding to the data which is being clocked out of the memory array, under the control of the processing unit. • Shunting the antenna of the Figure 2.7 causes a momentary fluctuation (dampening) of the carrier wave; this is seen as a slight change in amplitude of the carrier, as shown in Figure 2.8. However the equivalent circuit of rectifying unit in passive tag is shown in Figure 2.9.

Fig .2.7 Operating principle of a backscatter transponder. The impedance of the chip is 'modulated' by switching the chip's FET, [1]

13

CHAPTER 2

RFID SYSTEM AND THE COLLISION PROBLEM

Fig2.8 Amplitude- Modulated Backscattering Signal, [10]

Fig.2.9 Equivalent circuit of rectifying unit in passive tag. [20]

• The reader peak-detects the amplitude-modulated data and processes the resulting bit stream according to the encoding and data modulation methods used. [10]

2.5 Data transfer procedures – Full (FDX) and half (HDX) duplex procedures: transfer of energy from reader to tag is continuous and independent of data flow. Because the transponder's signal to the receiver antenna can be extremely weak in comparison with the signal from the reader itself, appropriate transmission procedures must be employed to differentiate the transponder's signal from that of the reader. – Sequential (SEQ) procedures: transfer of energy from reader to tag takes place for a limited period of time (power-supply pulses) and Data transfer from tag to reader takes place between power-supply pulses.[1] Figure 2.10 represents the data transfer procedures in RFID systems.

14

CHAPTER 2

RFID SYSTEM AND THE COLLISION PROBLEM

Fig. 2.10 Representation of full duplex, half duplex and sequential systems over time. [1]

2.6 Manchester Coding Manchester coding is usually adapted for collision detection in the signals that tag returns to reader. It enables the reader to detect the collision occurrence and the collision bit position. ‘1’ is represented by a high to low level change at midclock. ‘0’ is represented by a low to high level change at midclock. Manchester coding `10' and `01' are used instead of NRZ coding "1" and "0" respectively.

2.7 RFID frequency bands RFID systems typically operate within the freely usable ISM (Industrial, Scientific, Medical) frequency bands. Some of these bands, such as the 13.56 MHz ISM band, admit a single frequency channel for communication while other bands, such as the 915 MHz ISM band (in the US), admit multiple frequency channels for communication. However, most operating regulations for devices operating in ISM bands, such as FCC (Federal Communications Commission) Part 15 in the US, prohibit the explicit control of which channels are used for communication. [28]

15

CHAPTER 2

RFID SYSTEM AND THE COLLISION PROBLEM

The allowed frequency band for the RFID operation is limited. Hence the amount of taken frequency band and amount of exchanged traffic are important factors in comparing the proposed anti-collision protocols. The read range depends on the operating frequency band as shown in Table 2.3. Table.2.3 Read range against frequency range

RF Systems

Frequency Range

Typical Read Range

LF System

< 135 KHz

< 0.5 m

HF System

< 13.56 MHz

1m

UHF System

< 860 – 930 MHz

4-5 m

Microwave System

< 2.45 GHZ

1m

2.8 Reasons for the increased attention of RFID 2.8.1 Motivation for FRID RFID systems offer a promisingly affordable, cheap and flexible solution in many applications for object identification, including security and access control, transportation and automatic supply chain management and tracking, automatic inventory control. Inventories once taken by hand will be conducted and updated automatically. Continual database updates will better reflect the real world. At any moment, any item in the warehouse could be automatically located. [4] It is a technology that works well for collecting multiple pieces of data on items for tracking and counting purposes in a cooperative environment. For instance, fast and reliable reading of labels (tags) attached to different objects stowed in warehouses can greatly speed up operations such as localization and retrieval. Among the countless applications, we can cite public transportation and ticketing, access control, production control,

16

CHAPTER 2

RFID SYSTEM AND THE COLLISION PROBLEM

animal identification, inventory control, electronic payments, airline baggage, speed-up the checkout in shops, location monitoring in real time of all tagged objects within the supply chain, secure operations in dangerous environments, localization of objects (like cars in parking lots, or books in libraries), and of people. Due to advances in silicon manufacturing technology, RFID costs have dropped significantly. Low-cost RFID “electronic product codes” or “smart-labels” may be a practical updated replacement for optical barcodes on consumer items. The tag ID may be read automatically: without line of sight, through non-conducting material, at a rate of several hundred reads per second and from a distance of several meters. One day it is conceivable that every man-made object will be labeled with a unique identity associated with a digital entity. An object’s history, ownership, or location may all be available online.

2.8.2 Advantages of RFID over other auto-identification techniques RFID’s basic advantage over other identification techniques is the full automation of the data capture process where the optical identification systems fail. The most commonly used identification systems is the barcode system. Barcode systems typically require the laser gun to be shooted on the barcode to read it thus expecting an orientation between the two. The information in the barcode is fixed. The barcode system is sensitive to the clear optics, harsh environments and abrasion of the barcode on the item. RFID systems can work from a greater distance, even in harsh environments, without any need of the line of sight. The RFID systems read rate is about 50 tags/second in high frequency tags and up to

17

CHAPTER 2

RFID SYSTEM AND THE COLLISION PROBLEM

200 tags/second in ultra high frequency tags, which is very high compared to the barcode read rate. RFID systems can thus enable the tracking of items in real time. [32] * It does not require the tag to be in line-of sight. * Can work in harsh environments. * Multiple tags can be read simultaneously and automatically without human errors. * Variety of applications for RFID in manufacture, supply chain, retail inventory control, transportation, * tags can turn everyday objects into mobile network nodes, and then track, trace, monitor, trigger events, and perform some actions on those tagged objects.

2.9 Examples of recent RFID applications [42] Common applications range from highway toll collection, supply chain management, public transportation, controlling building access, animal tracking, developing smart home appliances and remote keyless entry for automobiles to locating children. In addition, RFID technology also offers a viable approach to implement physical user interfaces. The services available in the local environment are advertised by RFID tags. Users browse the services and activate the desired service by simply touching the corresponding tag with a mobile terminal that is equipped with an RFID reader. In the near future, these user interfaces would introduce RFID tags into our everyday lives. The purpose of that section is to show a part of the more recent uses of the RFID systems to improve the operation in different fields such as supply chain, tracking, aerospace, automotive, defense, health care, logistics and manufacturing.

18

CHAPTER 2

RFID SYSTEM AND THE COLLISION PROBLEM

2.9.1 RFID Tagged Euro Banknotes The European Central Bank (ECB) has announced the inclusion of RFID Tags in all Euro notes above €20 from February 2007 to improve counterfeit detection to counter the laundering of Euro banknotes.

2.9.2 Airbus Signs Contract for High-Memory RFID Tags The aircraft maker plans to use the EPC Gen 2 RFID tags—which will have as much as 8 kilobytes of memory—to track thousands of repairable parts and components, as well as store data, such as information regarding a part's initial construction and maintenance for its new Airbus' new A350 extra-wide body (XWB) fleet, expected to enter service beginning in 2013. The high-memory tags, to be placed primarily on repairable parts, will enable Airbus, aircraft owners and aircraft repair companies to improve their processes, such as maintenance and warehouse logistics. The FLYtag can be utilized on metal parts. The tag read ranges and the shape will vary based on the environment of use.

2.9.3Automotive Commercial truck and bus manufacturer owned by the Fiat Group plans to expand the RFID system it uses to process the receipt, picking and shipping of replacement parts, as well as guarantee their authenticity.

2.9.4 System for Tracking Health-care Assets A growing number of hospitals have deployed a real-time location system (RTLS) to track assets, patients and personnel. It is a low-cost alternative to RTLS, which generally employs battery-powered active RFID, ultrasound or infrared ID tags. The Asset Tracking solution, which utilizes passive EPC Gen 2 tags, would be quick to install, the companies report, and would allow hospitals to know when assets pass through or approach doorways.

19

CHAPTER 2

RFID SYSTEM AND THE COLLISION PROBLEM

The system could trigger alarms, automatically lock doors or transmit alerts to designated staff members.

2.9.5 Italian Construction Firm Deploys RFID to Track Offshore Equipment It attaches Omni-ID EPC Gen 2 passive tags to cranes, drilling rigs and thousands of other items, to improve safety and reduce wastage and delays. An Italian engineering and construction company for subsea oil and gas production has been tracking its large equipment on production sites for tracking 20,000 items, including offshore vessels (used to access oil drilling site), as well as cranes, drilling rigs, steel pipe, slings, shackles and buoys. Approximately 1,000 items have been tagged to date.

2.9.6 RFID Contains Solution to Shipping Problems For years, China International Marine Containers (CIMC's) inventorytracking process was very labor-intensive. Workers used a mix of optical character recognition (OCR) technology, paper, pens, and even binoculars in the container yards, to determine its products' whereabouts.

The

company's products are all made-to-order, and while they may look similar, each has its own unique features and specific weights, depending on customer requirements. As such, keeping track of each one is critical. Before the RFID pilot, when the containers were ready for transportation from the factory to the storage yard, truck drivers had to check out at the gate with piles of paperwork. This created backups, with drivers forced to wait in line until their papers were processed. This system caused inefficiencies and waste—in fact, the company often did not know the exact location of its containers, and in some cases, it lost them or delivered the wrong ones to its customers.

20

CHAPTER 2

RFID SYSTEM AND THE COLLISION PROBLEM

Fig.2.11 Real-time locating system (RTLS) in Chinese shipping system: Containers equipped with the passive tags for location-tracking purposes

In an effort to cut costs and improve operations, CIMC launched an RFID project to track containers from the factory to the storage yard, speed up the efficiency of global supply-chain management and reduce the cost of asset management. The technology could help lower costs and provide better reliability than OCR. Containers have been equipped with the passive tags for location-tracking purposes. Much of the work of accounting for the containers has been automated, with interrogators installed at the factory and storage-yard gates and on the forklifts used in the storage yard. Location information is transmitted in real time from a code division multiple access (CDMA) network to CIMC's existing yard-management system. This allows CIMC to know where its containers are at all times (Figure 2.11).

2.9.7 RFID Chips in Passports The U.S. Electronic Passport (e-passport) is the same as a regular passport with the addition of a small 64-kilobyte contactless integrated circuit (computer chip) embedded in the back cover. The chip securely stores the same data visually displayed on the photo page of the passport, and 21

CHAPTER 2

RFID SYSTEM AND THE COLLISION PROBLEM

additionally includes a digital photograph. The chip will contain the name, nationality, gender, date of birth, and place of birth of the passport holder, as well as a digitized photograph of that person.The inclusion of the digital photograph enables biometric comparison, through the use of facial recognition technology, at international borders The chip's contents will match the data on the paper portion of the passport, improving passport security by making it more difficult for criminals to tamper with passports. U.S. government efforts to make passports harder to forge began in response to the terrorist attacks on the United States on September 11, 2001. It was taking several security precautions. The RFID chips will use encrypted digital signatures to prevent tampering; and they will be so-called passive RFID chips, which do not broadcast personal information unless within inches of an RFID reader machine. To protect against data leaks, the e-passports will come with an "antiskimming" material that blocks radio waves on the passport's back and spine.

2.9.8 Payment by mobile phones Credit card companies are now looking for payment solutions for adding contactless payment cards to any mobile phone. Since summer 2009, two credit card companies have been working with Dallas, Texas, based DeviceFidelity to develop specialized micro cards. When inserted into a mobile phone, the microSD card can be both a passive tag and an RFID reader. A consumer can insert the card into his or her phone, then follow a few prompts on the phone's screen to enable it to link to the user's credit card or bank account, if so desired, in order to set up a payment system and used in mobile payment. Visa has announced plans to field test DeviceFidelity's In2Pay technology, a microSD-based NFC solution that enables mobile phones with a memory card slot to be used as a mobile payment device. 22

CHAPTER 2

RFID SYSTEM AND THE COLLISION PROBLEM

2.10 Problem of Data Collision in RFID system 2.10.1 Multi-reader Multi-tag Environment The operation of

RFID systems often involves a situation in which

numerous tags are present in the interrogation zone of a single reader at the same time. The tags can collide and cancel each other out, leading to retransmission of tag IDs that results in wastage of bandwidth and an increase in the total delay. Besides, readers physically located near one another may interfere with one another’s operation. Such reader collision must be minimized to ensure the current operation of the RFID system. [7] Collision problem can be divided into reader collisions and tag collisions. Anti-collision protocols for RFID system have been designed by taking into account that the powerful reader deals with little to no onboard computation capability (such as passive RFID tags). The main scope of this thesis is to survey the collision resolution protocols for the two collision problems and propose new protocols which give better and faster solution. 2.10.2 The Definition of Collision Problems in RFID In general, collision occurs when various devices interfere with each others operations, or their simultaneous operations lead to loss of data. The typical problem to be faced is a medium access control (MAC), i.e. to avoid or limit the number of transmission collisions. One of the main issues in RFID network is the fast and reliable identification of all tags in the reader range. The reader issues some queries, and tags properly answer. Then, the reader must identify the tags from such answers. Since the transmission medium is shared, both the readers and tags communicate over the same common broadcast wireless channel at the same frequency band. This can lead to the collision problem in signal transmission of the readers or the tags, which hardly leads to fast 23

CHAPTER 2

identification.

RFID SYSTEM AND THE COLLISION PROBLEM

As a result, either the reader may not recognize many

objects, or a tag identification process may suffer from long delay. A collision makes both the communication overhead and the transmission delay of a reader and tags have lost their useful ess. Therefore, anticollision protocols which enable the fast and correct identification regardless of the occurrence of collisions are required. The problem can be stated as: how to allow readers identifying tags with high data rate at the same shared channel within minimum time, although the very limited computation abilities of passive tags. 2.11 More challenging problem than multiple-access The reader queries the tags for their ID through the RF communication channel, by broadcasting a request message to tags. Upon receiving this message, all tags send an answer back to the reader. If only one tag answers, the reader identifies the tag. If more than one tag answers, their messages will collide on the RF communication channel, and the reader cannot identify these tags. This is a special case of the multiple-access communication problem. Since the low functional power and energy constraints in each passive tag (without internal battery), it requires the functionality of tag to be very simple, cheap and small enough so that its power consumption can be reduced. It is unreasonable to assume that tags can communicate with each other directly or that they can notice their neighboring tags or detect collisions. This places the full responsibility of detecting collisions on the reader. [4, 5] Hence, RFID introduce a more challenging aspect to the problem. 2.11.1 Constrains on anti-collision protocols in RFID systems 1- Lack of internal power source in the passive tags. This requires the tag reader to power-up these tags whenever it needs to communicate with them. 2- Total number of tags is unknown. 3- Tags cannot communicate 24

CHAPTER 2

RFID SYSTEM AND THE COLLISION PROBLEM

with each other. Hence collision resolution needs to be done at the tag reader. The low-functional passive tags cannot figure out neighboring tags or detect collisions. 4- Limited memory and computational capabilities at the tag. Thus the resolution protocol must be simple and incur minimum overhead from the tag’s perspective. These constrains are due to the requirement to keep the tags as cheap as Possible.

2.12 Types of data collisions in RFID system 2.12.1 Tags collision

Fig2.12 Collision due to tag multiple access problem. [11]

When multiple tags are simultaneously responding to the reader command on the shared wireless channel, reader will receive mixture of tags signals, so simultaneous responses from various tags prevent the reader from interpreting the signal correctly causing tag collision, which decreases throughput (Figure 2.12). Anti-collision algorithms are very important to solve the collision problem in multi-tag environment. Tag collision problem is more critical than reader collision because readers are assumed to have high functionality. While passive tags are mainly passive entities; they do not have enough power to differentiate between frequency ranges of the existing readers. Passive tags are constrained in terms of energy since they derive power from the readers signal only. Such a low energy supply requires the functionality of tag to be very simple so that its power consumption can be reduced. Hence collision resolution protocols for such kind of passive RFID tags should be simple enough. [6] 25

CHAPTER 2

RFID SYSTEM AND THE COLLISION PROBLEM

Since low-functional passive tags can neither detect collisions nor figure out neighboring tags, tag anti-collision protocols for passive RFID tags are important for fast and reliable tag identification. Manchester code is used, to enable the reader to detect where the collision bit is. As shown in figure 2.13, the zero signal will be encoded as one to zero transition, and the one signal as zero to one transition. The combination of the two signals equals two successive ones, which means existence of the two signals (i.e. collision state).

Fig .2.13 Response of tags with Manchester code. [43]

2.12. 2 Readers collision

Fig .2.14 Three intersected readers

Tags harvest energy from the reader communication field and use this energy to power any on-tag computations and communication to the reader. The limited power that can be harvested from the reader's signal severely

26

CHAPTER 2

RFID SYSTEM AND THE COLLISION PROBLEM

constrains the on-tag functionality and communication capabilities. Consequently, the reader-tag communication range is limited to a few meters or less. [28] There is asymmetric channel strength between the forward (tag-to-reader) and backward (reader-to-tag) channel. Hence, each reader has a limited reading range as shown in Figure 2.14. In some applications, for improving reading rate and ensuring the coverage the whole existing tags in the environment, several readers are put together in a network to form a dense RFID reader environment with intersected interrogation zones. Besides the existence of mobile RFID as shown in Figure 2.15, readers can be installed in a cellular phone and services are provided over telecommunication network. Readers collisions also result in reduction of the overall read rate of the RFID system. Hence reducing these readers collisions is essential. Moreover this problem is aggravated in case of mobile/handheld readers.

Fig.2.15 Mobile RFID reader

There are two types of readers collisions: as shown in Figure 2.15 (1) Multiple readers-to-tag interference occurs when the tag lies in the overlapped area of the reading ranges of several readers. More than one reader attempts to read with the tag simultaneously. The tag can not communicate correctly because it cannot distinguish the readers with

27

CHAPTER 2

RFID SYSTEM AND THE COLLISION PROBLEM

which it actually communicates. The tag hears multiple queries at the same time as shown in Figure 2.16.a. (2) Reader-to-reader interference occurs when there is unwanted transmission from nearby readers interfering with tags ability to decode a desired signal or masking the low level tag transmission from being recognized by the nearby reader, i.e. it arises when stronger signal from a reader interfere with the weak reflected signal from a tag. It can occur even without intersection of the reading zones as shown in Figure 2.16.b. Both reader and tags collision problems resolving will be discussed in details in the next two chapters.

Fig.2.16 Types of reader collisions

2.13 Performance evaluation of the anti-collision algorithm To judge an RFID anti-collision algorithm so that the tags can be read as fast and reliable as possible, the following criteria must be tested: (1) The number of transferred bits between the reader and the tags during the identification process. (2) Identification speed. (3) The required bandwidth. (4) The complexity of the algorithm. (5) The complexity of the logic operations in the tag side. (6) The possibility of working with the moving tags or readers. • Of course, fast algorithm means lower number of transmitted bits during the identification process. The complexity of the algorithm and the number of

28

CHAPTER 2

RFID SYSTEM AND THE COLLISION PROBLEM

consumed time slots to identify all tags are greatly related to the number of transferred bits between the reader and the tags.

2.14 The desirable characteristics of the anti-collision protocol To build an effective RFID anti-collision protocol, it must satisfy the listed characteristics in table 2.4. Table 2.4 the desirable characteristics of the anti-collision protocol. [25] Characteristics Minimal Delay Power consumption Reliability and Completeness Line-of-sight Independence Robustness Scalability

Description Time taken for identification of all the tags should be low. From a user point of view, this should not be perceptible. Due to the absence of an internal power source, power consumed by the tag should be minimal. The amount of power consumed is influenced by the total number of replies sent by the tag. An efficient protocol will minimize the messages between the tag and reader. All the tags in the range of the reader should get identified completely and correctly. The object attached with the tag can be located anywhere as long as they are in the range of the tag reading device (Reader). The protocol should work irrespective of environmental conditions. The protocol should be scalable to accommodate an increase in the number of tags.

2.15 Main system assumptions in building RFID anti-collision protocol Most assumptions made in classical networking cannot be similarly made in RFID systems due to some severe limitations such as the constraints on memory and computation capabilities, power limitations, and the inability to sense the medium condition. • The general assumptions in building anti-collision algorithms can be summarized as follows: (1) The reader is a powerful entity with abundant memory and computation power. (2) The tags are limited in memory and computation power. (3) There is a single shared communication channel between the reader and the tags. (4) The tags are not able to exchange messages among each other. (5) Upon receiving a message, each tag can optionally send a response back to the reader. (6) The reader detects the collision bit position on the channel by Manchester code. 29

CHAPTER 2

RFID SYSTEM AND THE COLLISION PROBLEM

In the next two chapters we will present a survey for the tags and readers anti-collision protocols respectively.

30

CHAPTER 3 TAG ANTI-COLLISION PROTOCOLS

Abstract The purpose of this chapter is to provide a survey of tag anticollision protocols. Large number of recent tag anti-collision protocols in literatures will be discussed. Moreover, this chapter presents the main classifications of anti-collision protocols for RFID systems. The advantages and disadvantages of each method will be discussed.

CHAPTER 3

TAG ANTI-COLLISION PROTOCOLS

CHAPTER 3 TAG ANTI-COLLISION PROTOCOLS 3.1 Introduction Tag collision is considered one of the great challenges in the RFID systems development. The RFID signal collision delays the tag recognition and causes loss of information. The incidence of multiple responses from multiple tags reaching simultaneously to the interrogator prevents it of identifying each response individually, unless some strategy can make the responses come in isolated time or anyhow controlled by the interrogator. The needed anti-collision protocols must be able to detect all tags presented in the reading area, although the collision occurrence, with minimal delay, minimal power consumption, reliability & completeness, line-of sight independence, robustness, and scalability. These protocols need to make a fast and correct identification of all tags present in a certain environment, just because they might get out of the reading area before being completely identified. [3] This chapter presents a survey of tag anti-collision protocols.

3.2 Main classification of tag anti-collision protocols

Fig 3.1 the main tag anti-collision protocol classification

30

CHAPTER 3

TAG ANTI-COLLISION PROTOCOLS

RFID tag anti-collision protocols proposed up to now to solve the tag collisions can be grouped as shown in Figure 3.1 into tree-based deterministic protocols and aloha-based probabilistic protocols. Researches on anti-collision protocols have been required to improve the characteristics of reading rate and the identification speed. This thesis is interested in developing an efficient deterministic binary tree protocol.

3.3 Probabilistic Aloha based protocols: ( ISO 18000-6A standard) [11] The stochastic model is often based on an ALOHA-like protocol. Aloha is a random access protocol used in communication networks to interconnect multiple radio terminals. Each tag sends their data at a random time period. When collisions occurred, the tags that are involved in collisions wait a random time with some probability before re-transmitting their ID again in the next read cycle. Technological advancements permit a version of Aloha called Slotted Aloha, in which the transmissions of signals are synchronized at the beginning of a slot. Each terminal waits for the available slot and transmits with a random probability. A slot is a time frame with limited number of bits; [12].

Fig.3.2 the basic slot aloha-based probabilistic algorithm. [11]

31

CHAPTER 3

TAG ANTI-COLLISION PROTOCOLS

In Aloha and Slotted Aloha, a tag responds after a random delay, and continues doing so until it is identified. In Slotted Aloha, a tag replies in synchronized slots. Because of the fixed frame size, the algorithm is inefficient. Most of time slots will be filled with more than one tag. The principals of the slotted Aloha protocol are shown in Figure 3.2. However, for Framed Slotted Aloha (FSA) and its variants, a tag selects a slot randomly and replies once in a frame. If there is a collision, tags defer to the next frame. Each tag transmits its serial number to the reader in a randomly selected slot of a frame, and the reader identifies tags when a time slot is used by one tag only. Read cycles are repeated for unrecognized (namely, whose transmission results in a collision) tags, until all tags have been identified. So, in the last read cycle there must be no collision. The frame size may vary over time. It generates an adaptive frame size for next cycle according to collision condition of last cycle. It determines the optimum frame size by counting the empty, collision and correct read time slots. In generally, the stochastic aloha-based protocols can fast complete reading, but, it cannot perfectly prevent tag collisions because of the probabilistic procedure. It has “tag starvation” problem where there is still a probability of failing to read all tags in a limited time period. It is more efficient and faster than binary tree protocols (which will be discussed later) when the number of tags is not very large and the ID length is short.

3.3.1 Tree Slotted Aloha (TSA) [27] The behavior of the TSA protocol follows a tree structure. It takes benefits of both Aloha and binary tree protocols. All tags select a slot to transmit their ID by generating a random number. If there is a collision in a slot, the reader broadcasts the next identification request only to tags which

32

CHAPTER 3

TAG ANTI-COLLISION PROTOCOLS

collided in that slot. The basic idea of TSA identification protocol is to solve a collision as soon as it happens as shown in Figure 3.3.

Fig.3.3 Example of TSA protocol execution

In Framed Slotted Aloha protocols, two tags not colliding in a frame can collide in the next frame. In TSA, the above situation is avoided, since when a collision occurs in a slot, only the tags that generate such collision are queried in the next read cycle. Consider an RFID system consisting of a reader and a set of n passive tags. Each tag t

{0, ..., n − 1} has a unique ID string tid

{0, 1}k, where k is

the length of the ID strings. It assumes that the reader does not know the exact number n of tags present in its communication range, but it can estimate it. Such an estimation l0 is the starting frame size. The initial estimation does not affect the performance. The protocol is performed in several tag reading cycles. A reading cycle consists of two steps: in the first step, the reader broadcasts a request for data by specifying the frame size li, in the second step each tag in the communication range of the reader, selects its response slot by generating a random number in the range [1, ..., li] and transmits its ID in such a slot. The reader identifies a tag when it receives the tag ID without collisions. At the end of each reading cycle, if the reader realizes that collisions occurred, it starts a new reading cycle for each slot where there was a

33

CHAPTER 3

TAG ANTI-COLLISION PROTOCOLS

collision. This corresponds to adding new nodes in the tree, as sons of the node representing the above reading cycle, one son for each slot with collisions. The reader broadcasts such a size of new cycles, together with the slot number of the previous frame (to address only the tags colliding in that slot), and the level of the tree. In each reading cycle, tags store the generated random number (i.e. the slot in which they transmitted their ID) and increase by one their own tree level counter, so that they can realize when are involved in later communications. TSA is not memoryless, since each tag has to remember the random number generated in the previous cycle, and the level of the tree. The whole process is recursively repeated until no collisions are detected in a cycle. It performs better than Framed Slotted Aloha and query tree based protocols, in terms of number of slots needed to identify all tags, which is a commonly used metric, strictly related to delay.

3.4 Deterministic Binary Tree Protocols The deterministic protocols identify tags by constructing binary trees through the binary bits of tag IDs. The deterministic binary tree anticollision protocol explores tags based on their unique identification numbers. It walks or traverses the tree all the way down until it identifies the tag. The unique tag number is determined from the nodes of the tree; [12]. Deterministic resolution of collisions based on binary tree algorithms is done by successively splitting tags into two subsets, directs tags to either remain active, or go temporarily inactive. Only one tag will be left sending its data at last.

Such deterministic methods can be classified into a

memory based algorithms and a memoryless based algorithms.

34

CHAPTER 3

TAG ANTI-COLLISION PROTOCOLS

In memory algorithms, the response of a tag is determined by the query to the tag and the current state of the tag. Thus each tag must remember its state. In memoryless algorithms, the current response of a tag is determined only by the current query of the reader to the tag. Memoryless algorithm has advantages of simple tag implementation as well as low cost. Moreover, the only computation required for each tag is to match its ID against the binary string in the query. There is a trade off between simplicity and speed. Although tree-based protocols do not cause tag starvation, they have relatively long identification delay due to the splitting procedure starting from one set including all tags; [21].

3.4.1 Query Tree (QT) Protocols It is a memoryless tag identification protocol, in which the tags do not need to remember their inquiring history. Hence, it is requiring very simple tag circuitry. The QT algorithm consists of rounds of queries and responses. In each round, the reader asks the tags whether any of their IDs contains a certain prefix. That section presents different versions of query tree protocols. 3.4.1.1 Basic Query Tree Scheme The reader sends a query containing a prefix having length of 1 to n bits. The tags whose prefixes match with the bits sent by the reader, replies back with their tag ID. [6] The reader asks the tags to answer if their ID matches a given prefix. If there is a collision, the reader queries for one bit longer prefix until no collision occurs. Once a tag is identified, the reader starts a new round of queries with another prefix. An example is shown in Table 3.1 where there are 3 tags with the IDs of {0000, 0010, 1000}. 35

CHAPTER 3

TAG ANTI-COLLISION PROTOCOLS

Table 3.1 the tag identification process of Query Tree Protocol (*: collision)

Reader sends Start 0 1 00 01 000 001

Tags answer

Status

* * 1000 *

Collision Collision Read Collision No response Read Read

0000 0010

Queue States (Initially {}) {0,1} {1,00,01} {00,01} {01,000,001} {000,001} {001} {}

A queue of such prefixes is maintained and the queries are sent in order from this queue. As and when a query is done, its corresponding entry is removed from the queue. If there is a collision corresponding to any query, i.e., there are more than one tags with the same prefix, the reader removes that query and adds two new queries to the queue, first by appending a 0, and then a 1 to the current prefix. Therefore, by extending the prefixes until only one tag's ID matches. No collision implies either an ID has been read successfully or there is no tag with matching prefix. 3.4.1.2 Query Tree with Reversed IDs (QTR) [13] This protocol works by reversing the IDs of the tags and then applying the query tree (QT) protocol. QTR protocol outperforms QT protocol if the tags IDs are consecutive integers. The starting idea is that, when classifying the several numbers of different bit strings, if the bit string has consecutive or identical prefix, it is effective to classify the suffix first. From this basis, the ID of most tags has identical and consecutive prefix from manufacture to retail store on the progress of supply chain management.

36

CHAPTER 3

TAG ANTI-COLLISION PROTOCOLS

3.4.1.3 Query Tree-Based Reservation (RN16QTA) [14] It starts by generating short16-bit random temporary IDS (RN16s) by all tags in the field of a reader for giving the uniqueness to themselves, then applying the Query Tree Algorithm(QTA) with the temporary ID, first to be identified through its virtual ID, then to send its real ID. • Request: The reader sends n-length inquiring bits (prefix) to tags. • Response: Tags send their RN16s from (n + 1) th bit to the end when the first n bits of the tag IDs are the same as the prefix. • If the reader identifies a RN16, the reader calls the tag having the RN16 with an ACK command. Figure 3.4 shows the types of time slots. • Only the tag sending the whole RN16 is activated by the ACK command of the reader, and responds with its real tag ID to the ACK command.

Fig 3.4 successful, collided, empty time slots of RN16QTA 37

CHAPTER 3

TAG ANTI-COLLISION PROTOCOLS

•The RN16QTA reduces the overhead from both the average inquiring bits and response bits by using the temporary ID. It performs the query tree anti-collision algorithm on a smaller length of 16-bit randomly generated (virtual or temporary ID) numbers as shortcut representation of the original ID with a more length. It tries to identify tags through identifying their randomly generated numbers first, then request the tag that own that virtual ID to start transmitting its full length real ID . It minimizes the problem of identification of full length ID to a problem of identifying colliding tags of randomly generated 16-bit IDS. •The overhead of the identification process is the exchanged traffic for identification of the 16-bit length virtual IDS. 3.4.1.4 An Adaptive Memoryless Tag Anti-Collision Protocol for RFID Networks [15] It improves the performance of query tree protocol by maintaining the history of read cycles in the form of Candidate Queue (CQ). Candidate queue maintains the list of leaf nodes of the last query process. To shorten the collision period, it uses information which is acquired during the last identification process. The reader maintains not only the queue Q but also a candidate queue CQ. Before the process of identification, the reader initializes the queue Q with the candidate queue CQ which stores prefixes from the last process of identification. The candidate queue CQ manages prefixes with the query insertion procedure and the query deletion procedure.

3.4.1.5 Intelligent Query Tree Protocol (IQT) [6] IQT exploits the existence of the common prefix patterns in the tag IDs to reduce the communication overhead between the reader and tags.

38

CHAPTER 3

TAG ANTI-COLLISION PROTOCOLS

The common prefix in tag IDs may be due to the fact that items to which the tags are attached have the same manufacturer or product type as shown in Figure 3.5. Query Tree Protocol (QT) has been modified for the scenarios where the tags within the range of the reader have some common prefix. It also uses history of read cycles to further improve the tag read efficiency. Common prefix area (Overlapping Region)

Fig.3.5 IQT protocol achieves saving common prefix bits, [6]

When it performs subsequent read cycles, it knows whether the tags have some common prefix. If they have a common prefix, then it is known to the reader. Hence, IQT does not read those bits again. Hence, if all the items have first prefix bits common, then communication can take place using only the remaining (rem) bits IQT Protocol achieves substantially better performance with very minimal change in tag hardware complexity. This is because along with the prefix, the query itself can contain control information to determine whether query will take place using the entire tag ID or with the remaining bits. Better efficiency can be obtained by decreasing the number of bits transmitted between reader and tag in one query, as well as by reducing the number of queries in one read cycle. 39

CHAPTER 3

TAG ANTI-COLLISION PROTOCOLS

Following are its advantages over Query Tree Protocol: • Reduction in number of bits transmitted by tag. • Reduction in number of collisions by maintaining the history of tag read patterns. • Reduction in number of collisions for subsequent read cycles by using the information of previous read cycles.

3.4.1.6 The collision tracking tree algorithm (CTTA) [25] CTTA is based on Query Tree Algorithm (QTA) except that this scheme uses collision tracking. The difference between CTTA and QTA is as follows: Tag: the tags send their IDs from (k+1)th bit to the end bit if the prefix is the same as the first k bits of tag IDs. However, the tags stop sending their IDs when an ACK signal is received. Reader: the reader checks whether a collision occurs or not in each bit on the received sequences, and transmits an ACK signal to stop being sent the tag IDs by the tags if there is a collision. If there is a collision at nth bit in the received sequences, the two new prefixes, ‘the former prefix k bits + the received n-1 bits + 0 or 1’, are retransmitted sequentially to the tags in the field of the reader. Furthermore, if there is no collision, the reader identifies a tag corresponding to the detected ID, which is the connection of the prefix and the response. RFID systems do not have to use the whole length of tag IDs for resolving each tag in the field of a reader. According to the characteristic of QTA and CTTA, QTA requires significantly more bits for both the average inquiring bits and response bits for one-tag identification than CTTA.

40

CHAPTER 3

TAG ANTI-COLLISION PROTOCOLS

3.4.1.7 Anti-collision Algorithm based on Binary Tree Searching Algorithms and Balance Incomplete Block Design BIBD (4, 2, 1) [16] In this scheme a tag is divided into several sections, and each section contains special sequences. Readers can get every ID by identifying each section one by one. Each section consists of two bits which were represented according to Manchester Coding. In Manchester coding `10' and `01' are used instead of NRZ coding '1' and '0' respectively. Because in Manchester coding there is no `00' and `11' unites, the detection result is`0xx1' which has two collision bits. ** The existing subsets can be get from collision sequence because BIBD(4,2,1) only includes 6 subsets {0011,0101,0110,1001,1010,1100}: * Two collision bits in back sequence (received from tags): there are 2 subsets whose collision bits are 01 and 10 from left bit to right. For example from 0x1x, 0011 and 0110 can be identified. * Three collision bits in back sequence: there are 3 subsets in this sequence. When the no-collision bit is 0, the collision bits are 011, 101 and 110 from the highest bit to the lowest respective. When the no-collision bit is 1, the collision bits are 001, 010 and 100. For example from 0xxx, 0011, 0101 and 0110 can be identified. * Four collision bits in back sequence: There are 2 ‫ ـــ‬4 subsets in this sequence. Round identification method is used. The searching speed of the algorithm is six times that of the Query Tree Protocol. It is especially suitable for the identification conditions which contain numerous tags or tags with long ID. However, due to multi-section processing, the identification time of this method is relatively high.

41

CHAPTER 3

TAG ANTI-COLLISION PROTOCOLS

3.4.2 Binary Tree Anti-Collision Protocols In general, binary tree anti-collision protocol, shown in Figure 3.6, identifies tags based on their unique identification numbers. Inside each tag, there is a pointer. Every time the tag is reset and the pointer points to the highest bit of the tag's ID.

Fig.3.6 Binary tree representing the EPC number

During the reader inquiry, it sends one inquiring bit at a time. The tags whose pointed bit is the same as the inquiring bit will back send their next bits to the reader, and the tags whose pointed bit is not matched, will convert to the state of “standby” and will not answer the remaining inquires in this round. This ongoing of inquiring moves toward the lowest bit. This process will be continuing until one tag has been identified and then all the remaining tags are reset. When the reader senses a non-collision answer it uses the next-step inquiring bit. But if a collision is sensed it uses a ‘0’ bit as next-step of inquiring. Thus every cycle of inquire only one tag will be identified when the pointer reaches to the lowest bit of the tag's ID. Once a tag has been completely identified, it will be eliminated. Then all the other tags that have already entered the state of “standby” will be reset. A new cycle of inquiring again start from the highest bit. After k cycles of inquiring, the IDs in the k tags will be identified. This section presents different versions of binary tree anti-collision protocol. 42

CHAPTER 3

TAG ANTI-COLLISION PROTOCOLS

3.4.2.1 Dynamic Bit Arbitration (DBA) Anti-Collision Algorithm for RFID System [17] It aims at reducing the number of transferred bits between the reader and the tags, because the speed of identification process and the energy consumption are greatly related to the number of transferred bits. Table 3.2 the anti-collision process of DBA [17] Reader Reader steps Communication Register

 

Sending the restart point, after each tag identification

1 2  3  4  5  6  7  8  9  10  11  12  13  14  15  16  17  18  19  20 

Request All 0 0 1 0 1 1 0 Request(5) 1 0 1 1 Request(7) 1 1 0 1 1 0

7 7 7,5 7,5 7,5 7,5 7 7 7 7 7

Tag Answer (Register) Tag 1 Tag 2 Tag 3 Tag 4 00101101 00110110 00110111 01101101 0 0 1 0 1 1 0 1

0 0 1 1 (5) (5) (5) (5) 1 0 1 1 0

0 0 1 1 (5) (5) (5) (5) 1 0 1 1 1

0 1 (7) (7) (7) (7) (7) (7) (7) (7) (7) (7) (7) 1 1 0 1 1 0 1

If the collision occurs, the reader will record the position of the collision (k) in its register. The reader will choose a path by replying with 0 or 1. Some of tags, which just sent the same response value to the reader continue the next bit arbitration step, while the others record the stopped position in their registers and enter the Quiet state at the same time. Then it returns to the nearest collision bit by checking its register and transmitting request command including the nearest collision position. 43

CHAPTER 3

TAG ANTI-COLLISION PROTOCOLS

The tags with the same value in their registers quit the Quiet state and proceed with the bit arbitration step from where they stopped. The other tags still stay in the Quiet state until they receive the appropriate request command from the reader. The reader and the tags will repeat the above steps until all tags are successfully identified. However, if the collision occurs in the last bit of the tags' IDs, the reader can identify two tags simultaneously. Table 3.2 gives an example of four tags identification using DBA protocol.

However, this algorithm transmits overhead

information, after identifying each tag, to inform the tags in the nearest collision node to change its current Quiet state to active state. 3.4.2.2 Improvement to the Anti-collision Protocol Specification for 900 MHz Class 0 Radio Frequency Identification Tag [18]

Fig.3.7 Representation of an EPC as a tree.

This process is shown in Figure 3.7. It reduces the overall read time of a given number of RFID tags by resetting to the appropriate node, for every consecutive read cycle. The first few MSBs have the same pattern; hence, instead of always going to the root of the tree for every read cycle, the Tree Traversal cycle can begin with the bit position from where the last difference in the tag IDs occurred.

44

CHAPTER 3

TAG ANTI-COLLISION PROTOCOLS

It increases the complexity of the tag by including a conflict-counter and a conflict-bit pointer. While in the active state, when a tag conflicts at a certain bit, it records the bit position in the conflict-bit-pointer, increments its conflict-counter by one and moves to the Mute State. While in this state, it increments the conflict-counter every time it detects a conflict between the rest of the tags in the active state and decrements the counter for every successful read. When the tag receives a ‘Null’ bit after another tag has been successfully read, it transits to the Tree Start state.

3.4.2.3 ID-Binary Tree Stack Anti-collision Algorithm [19]

The need of sending query of last collision bit position after each identification to wake up inactive tags that was stopped at that bit position.

Fig.3.8 ID-binary tree stack anti-collision algorithm. It takes only 24 bit transmission between tags and the reader to identify the four tags

This technique is illustrated in Figure 3.8. It reduces the traffic between the reader and the tags. By storing the threads (branching points) of the identification in a stack, the only computation required for each tag is to 45

CHAPTER 3

TAG ANTI-COLLISION PROTOCOLS

count the stack depth. The tag uses a counter to keep track of the stack position, where the tag is on. Only the tags whose stack counter are zero can take part in the later communication with the reader. Advantage: each tag only transmits every bit of the ID once in the whole collision arbitration process. The tags communication complexity is equal to the length of the ID. The cost is that each tag needs an additional counter which works as a pointer to the stack. The tag uses its internal counter to determine when it changes its state from Quiet state to active state. But, the tags in active state need to know the last collision position to restart from. In this thesis, we aim to further reduce the traffic between reader and the tags by deleting query transmission of last collision bit position after each identification to inform the waking up tags with the restarting bit position.

3.4.2.4 An Efficient Tree-Based Tag Anti-Collision HeightOriented Protocol [20] This protocol aims to minimize the total time complexity to identify all tags in the interrogation zone, and to minimize the power consumption of tags, which results in a larger interrogation zone, by achieving faster and simple logic anti-collision algorithm. The basic search criterion is based on the depth-first search (DFS) algorithm. A height-oriented method of using an encoded binary number corresponding to the height of the vertex at the tag collision most recently occurred, which leads to prevent the redundant increase of the transmission bit length.

Each tag has a pointer, where the pointer moves toward a

lower bit with the ongoing of inquiring, and a reader has the function of recording the position of the vertex at which the tag collision occurred. It resolves the disadvantage of conventional method of applying a prefix binary number from the root to a vertex, which causes the redundant

46

CHAPTER 3

TAG ANTI-COLLISION PROTOCOLS

increase of the transmission bit length. Figure 3.9 to Figure 3.11 demonstrate this protocol. Relatively long reader command frame per one bit tag response.

Fig.3.9 Bit stream between a reader and tags in the interrogation zone, and the flow diagram of searching tags.

Fig.3.10 State diagram of the protocol.

47

CHAPTER 3

TAG ANTI-COLLISION PROTOCOLS

Fig.3.11 Tag ID pointer

Disadvantages:

the main disadvantage of this protocol is the large

overhead due to: (1) The need to send the height of the most recently occurred collision bit position after each identification to wake up inactive tags that was stopped at that bit position. (2) The reader command frame is relatively long per each transmitted tag bit. Minimizing the transmitted overhead problem is the main target of anticollision protocols.

In the proposed Parallel Binary Splitting (PBS)

protocol in the next chapter, we aim to further reduce the traffic between reader and tags, by deleting query transmission of last collision bit position after each identification to wake up inactive tags that was stopped at that bit position.

3.4.2.5 Adaptive Binary Splitting (ABS) Tag Anti-Collision protocol [21], [22] ABS is an improvement of the traditional binary tree protocol. ABS protocol improves ISO 18000 6B protocol. ABS assigns distinct timeslots to tags by using obtained information from the last identification process. ABS protocol is a quite good idea, using counter to achieve branching and further achieve anti-collision effect. Each tag is provided with two counters to determine the timeslot for the transmission as follows: • PSC (progressed-slot counter): is used to store the number of recognized tags in the ongoing process. PSC is increased when reader feedbacks readable timeslot, and decreased, when feedbacks Idle timeslot. All the tags have the same value of PSC at all times. • ASC (allocated-slot counter): signifies the timeslot for transmission its ID. The tag is allowed to transmit when its ASC is equal to PSC. It

48

CHAPTER 3

TAG ANTI-COLLISION PROTOCOLS

assumes that the channel is slotted and a reader performs identification processes repeatedly for object tracking and monitoring. In a timeslot, tags transmit ID to the reader and the reader then sends feedback.

Fig.3.12 ABS anti-collision algorithm

This protocol uses random binary numbers generated by colliding tags for the splitting procedure. The tag has a counter initialized to 0 at the beginning of the frame. The tag transmits its ID when the counter value PSC=ASC=0. Therefore, all the tags within the reader’s reading range, at the start of the frame, form one set and transmit concurrently. The reader sends feedback to inform tags of the occurrence of collision. According to the reader’s feedback, all tags change their counter as shown in Figure 3.12. To control PSC and ASC, the reader informs tags the type of the last interrogation cycle, namely, idle (no tag signal), readable (only one tag signal) or collision (multiple signals) by sending feedback. According to the feedback, tags act as follows: • Readable (the reader recognizes a tag successfully): Tags add 1 to PSC. If the tag has ASC less than PSC, it does not attempt the transmission until the completion of the frame because it has been recognized in the ongoing

49

CHAPTER 3

TAG ANTI-COLLISION PROTOCOLS

frame. The tag recognized by a reader does not transmit any signal until the ongoing frame is terminated.

Fig.3.13 ABS tag operation: (a) collision cycle (b) idle cycle. [22]

• Idle (No transmission is attempted): Tag decreases ASC and PC by 1 as shown in Figure 3.13.b, to pull the schedule of the transmission if PSC < ASC.

• Collision: When a tag collision occurs, the tag which has ASC is equal to PSC randomly selects one of two binary numbers, 0 and 1, and then adds it to ASC to split a set of transmitting tags at the same interrogation cycle as shown in figure 3.13.a,. Note that colliding tags are the current active tags (i.e. ASC equal to PSC). * The active tag which is involved in this collision (i.e., the counter value is 0) randomly selects a binary number (0 or 1). By adding the selected binary number to its allocated slot counter, a set is split into two subsets. * The tag which is not involved in this collision (i.e., the counter value is greater than 0) increases its counter by 1. *Since PSC is not changed at collision timeslot, the first subset (tags which generate 0) stays active and retransmits at the following timeslot, while the second subset (tags which generate 1) retransmits after the first subset is recognized. To prevent the second subset(tags which select 1) and another

50

CHAPTER 3

TAG ANTI-COLLISION PROTOCOLS

set of tags, which have already had the allocated slot counter of ASC + 1, from integrating, tag adds 1 to ASC if PSC < ASC. Drawbacks of this protocol • ABS uses counter to achieve the goal of anti-collision, but the splitting of sets depends on the random binary number {0, 1}. So, it cannot provide the best splitting result. The probability of occurrence of 1 or 0 is not 50%.At any moment, there won’t be any splitting result, and may cause the next timeslot to be an idle timeslot or collided timeslot.[23] • The tag ID transmitting time is too long. Consumption of timeslots and longer timeslot are the main defects of ABS protocol. [23 ] • In collision, it performs splitting according to random binary counter that changes allocated slot counter, to make different reply order, before retransmitting the whole ID. **The main difference between this protocol (ABS) and our proposed protocol (PBS) is the consideration of the splitting rule. This difference will be discussed later.

3.4.2.6 An Enhanced Anti-collision Algorithm Based on Counter and Stack (EAA) [23] In this protocol, the main assumption is the reader ability of truncating unnecessary data bits to reduce the receiving time at collision condition. It means that reader does not need to receive any data after it receives the second collided bit. The reader shall determine this timeslot is a collided timeslot. It uses Manchester code, which is used in binary search algorithm, to find where the collision bit timeslot is. The process of EAA identification is illustrated in Figure 3.14.

51

CHAPTER 3

TAG ANTI-COLLISION PROTOCOLS

• The reader uses stack and Tstring to record the response received from tags. So, tag needs to reply the part of tag ID only in each timeslot. • Every tag (i) has an Ac(i) (Allocated-slot counter), a pointer(i), and a Pc (progressed-slot counter). Tag(i) replies to reader depending on Ac(i), pointer(i), and Pc. • When Pc(i)=Ac(i) and pointer is k, tag(i) shall reply its ID from kth bit to nth bit to reader.

Advantage The reader stop receiving after the second collided bit (Collided timeslot).

Advantage Restart in another identification path from the last branching node. It is better than restarting in the tree root

Fig.3.14 process of EAA identification example. [23]

This protocol defines three kinds of timeslots: 1. Collision: If there are more than two tags whose Ac(i) are equal to Pc in timeslot and the number of collided bit is larger than 2. 2. Identified: If there is only one tag reply in timeslot, then we call timeslot is an identified timeslot.

52

CHAPTER 3

TAG ANTI-COLLISION PROTOCOLS

3. OBCT (One Bit Collision Timeslot): If there are more than two tags whose Ac are equal to Pc in timeslot and the number of collided bit is equal to 1. • The operating codes are in 3bit; 000, 001, and

010 mean collision,

identified, and OBCT, respectively. Figure 3.14 shows an example to identify four tags (0000,0110,1110,1111). It consumes five timeslots. The total number of data bits which are exchanged between tags and reader is 19+11=30. The total number of bits of feedbacks and response which is received by reader is 19 and 11, respectively. It reduces not only the number of timeslots, but also the length of a timeslot. Owing to the cost to implement a counter in a tag being very low, it is valuable and feasible to get a better performance. • It achieves advance by transmitting a part of

ID starting from

collision bit to the end of the ID, so decreasing the length of the time slot. • The reader sends the position of the most recent collision bit, after each tag identification, to restart in another identification path from the last branching node. It is good to restart from that node; it is better than restarting in the tree root. But, it still wastes time in the transmission of the position of collision points. That time will be increased and largely amplified by increased collision due to increasing number of existing tags. The number of collisions is proportional to the number of existing tags. The main drawback of this protocol is the Long reader feedback message that must inform tags with the result of last timeslot information (OBCT: one bit collision timeslot ‫ ـــ‬Identified ‫ ــــ‬collision). In collision condition, it must feedbacks the position of the first collided bit. It represents the main data overhead of that protocol. 53

CHAPTER 3

TAG ANTI-COLLISION PROTOCOLS

3.4.2.7 Bi-Slotted Tree based Anti-Collision Protocols for Fast Tag Identification [24], [25] It proposes three methods for fast tag identification: bi-slotted tree based RFID tag anti collision protocols, query tree based reservation, and the combining method of them for enhancing the identification speed of RFID systems. Focusing on collision condition of the query tree based algorithms which send a prefix twice except the last bit in the same tree depth. The crux of these protocols is that an RFID reader transmits one ‘n − 1 length inquiring bits’ at a node which is related to a collided nth bit in the received tag IDs instead of two ‘n length inquiring bits’, which have the same first n − 1 bits and the different last bit. Then the corresponding responses allocated for tags

consist of two time slots depending on

whether nth bit is ‘0’ (first slot) or ‘1’ (second slot). This technique reduces the average required prefix overhead. It is diminishing prefix and iteration overhead to minimize the anti-collision cost for fast identification. It uses time divided responses depending on whether the collided bit is ‘0’ or ‘1’. The reader sends n-1 length inquiring bits (prefix) to tags. Matched tags respond their IDs to the reader, they choose one of two time slots depending on whether nth bit is ‘0’ (first slot) or ‘1’ (second slot). Thus, the time slot indicates the value of nth bit. *BSQTA (bi-slotted query tree algorithm): tags send their IDs from (n+1)th bit to the end bit. *BSCTTA (bi-slotted collision tracking tree algorithm): Tags send their IDs from (n+1)th bit to the time that ACK signal, which is sent from the reader when a collision occurs, is received. If a collision occurs at the last bit in tag IDs, the reader assumes there are two tags because of the uniqueness of the tag IDs.

54

CHAPTER 3

TAG ANTI-COLLISION PROTOCOLS

BSCTTA requires less average required bits for one-tag identification because of the reduced inquiring overhead. Figure 3.15 shows the Flow Chart of BSCTTA. Figure 3.16 shows the average required inquiring and response bits for one-tag identification. According to that figure, BSQTA reduces the average required inquiring overhead to ‘half - 1bit’ of the inquiring overhead in QTA without any increase in the average required response bits for one-tag identification. Thus, BSQTA requires less average required bits for one-tag identification than QTA.

Fig.3.15 Flow Chart of BSCTTA, [25]

Figure 3.17 shows that BSCTTA is the fastest identification protocol, because it can read more tags per second. To calculate the average number 55

CHAPTER 3

TAG ANTI-COLLISION PROTOCOLS

of identified tags per second, the RFID systems choose 8 bits for request and 3 bits for detecting collision, no collision, or no response for substituting the iteration overhead.

Fig.3.16 Search Cost (Bits): BSQTA and BSCTTA

Fig.3.17 Identification Performance: BSQTA and BSCTTA

56

CHAPTER 3

TAG ANTI-COLLISION PROTOCOLS

3.4.2.8 An Anti-Collision protocol of RFID Based on Divide and Conquer Algorithm (DCQT)[26] The divide and conquer algorithm divides a big problem of size n uniformly at into two small sub problems of size k and n−k that can be solved easily and quickly, repeating to the same solving steps to solve these small sub- problems, and then somehow combine their solutions to obtain a solution to the original problem. The algorithm divides the successive collision bits into several two bits groups, and uses these groups to search which the collision group is and where the collision occurring, and identify all tags finally. The character string is composed with character “0”, “1”, and “C”. Where“0”represents the logical 1, “1”represents the logical 0, and “C” indicates that has a collision at that bit. It needs to replace the collision bits with logical 0 or logical 1, and then it should show all the possible IDs and identify each of them. It divides the successive collision bits into several groups, and each group is two bits; as shown in Figure 3.18.

Fig.3.18 The Possible results of Groups of Collisions for the four Successive Collision bits.

The reader of the proposed anti- collision protocol algorithm decomposes the collision bits of the received ID into several groups first. Each group is two bit group; the whole groups are 00, 01, 10, and 11. Then, the reader search tags one by one to find which groups are the collision groups and where the collision occurring now. 57

CHAPTER 3

TAG ANTI-COLLISION PROTOCOLS

In the research step, the reader sends the possible collision groups and the collision positions to tags at the same time. When there is tags whose IDs match with the reader sent collision groups and collision positions, then those tags just do a response to the reader and don’t care the response contents. The response contents are not important for the reader, because the reader only wants to know which the collision group is and where the collision occurring. The worst environment for a reader is that there are 2^B in its receiving area (range) and collision occur at every bit. In the worst environment, there are B groups and each group has 4 possible different contents, so that the number of communications is 2B = B*4/2. The upper bound of the number of collisions is the number of communications, So that the number of tag collision is really no more than the two times of the tag length. The length of a standard RFID tag ID is 96 bits now, and to increase the length of the standard RFID tag ID is possible in the future. The worst condition for a reader now is that there are 2^96 IDs in its emitting & response receiving area, and the number of tag collision is really no more than 2*96 = 192 the number of tag collision depends on the length of tags, and is independent of the number of tags. Hence it leads to upper bound to the number of tags collision being resolved by reader queries. Table.3.3 Comparison of querying times of different algorithms.

58

CHAPTER 3

TAG ANTI-COLLISION PROTOCOLS

Table 3.3 give a comparison of the number of queries sent by the reader under the applying the anti-collision protocol of

query tree algorithm

(QT) , query tree reversed algorithm (QTR) and query tree based on divide and conquer (DCQT) to identify all tags completely. It is clear that, the querying times of QT and QTR are direct proportion to the number of tags. While the querying times of DCQT is independent of the number of tags and is a constant number, and has upper bound.

3.4.2.9 Novel Anti-Collision Algorithm for Identifying Passive Tags (NEAA) [43] A new idea in [43] is introduced to reduce the probability of collision efficiently and to make fast identification of multiple passive tags in a timeslot. It reduces the length of the time slot by truncating unnecessary data bits to minimize the receiving time. The reader does not need to receive any data after receiving the first collided bit. It uses Manchester code to locate the collided bits. For example, the IDs of tags A and B are “10011111” and “10101101”, respectively. Since tag A and tag B reply in the same timeslot, the combined response which was generated by tags A, and B is “10xx11x1” by the decoding method of Manchester code (“x” means a collided bit). It classifies the tags into sets according to the number of ones “1” in each tag-ID. Then it can identify tags set by set as shown in figure 3.19. For example, if the number of data-1 in the ID is 1, then the three IDs “001”, “010”, and “100” of tags A, B, C, respectively, can be identified in one time slot by the use of Manchester code. This algorithm uses four types of timeslots are used. These time slots are: collision, readable, M-readable (Multiple-readable), and idle.

59

CHAPTER 3

TAG ANTI-COLLISION PROTOCOLS

A feedback message is sent by the reader to inform the tags about the type of a timeslot. Based on this message, all tags can make a suitable response for the next timeslot. When collision occurs, the reader will inform all tags about the collided bit. Feedbacks are just like instructions and include operating code and some other information. The operating code in 3 bits is used. It improves the collision resolution, reduces the length of the time slot and makes upper bound of the total time slots is needed for identification. So, it is a long reader instructions and the reader must inform the tag the position of collided bit in the collision code.

Fig.3.19 NEAA tree shows process of NEAA step by step.

It gives an example of identifying seven tags which are denoted A, B, C, D, E, F, and G in the interrogating zone, and their tag IDs are “0000”, “0001”, “0010”, “0110”, “1001”, “1010”, and “1110”, respectively. So, the total number of data bits which are transmitted between tags and reader is 23+19=42. The total number of bits of feedbacks and total number of bits of responses are 23 and 19, respectively. 60

CHAPTER 3

TAG ANTI-COLLISION PROTOCOLS

However, the upper bound is computed relative to the number of time slots. It has four type time slots with three bits instruction code, besides sending the position of the collided bit along with the collision code.

3.4.3 Performance under moving scenario The relative movement between the reader and the tags leads to some tags are arriving or leaving the read range of the reader. The current questions: • Can the anti-collision protocol continue in the identification session in the case of moving tags or reader without disturbing the identification process? • Can the anti-collision protocol exploit information obtained from the last frame to prevent collisions between staying tags, and also avoid collision between arriving tags and staying tags?

3.4.3.1 Faster re-identification in successive interrogation sessions In many RFID applications, the reader may repeatedly identify existing tags. Thus, when there are a lot of tags which don't leave the reader's range, called staying tags, then if an anti-collision protocol can retain information obtained from the last process of tag identification, i.e., last frame, the reader can skip many collisions and quickly re-identify the staying tags in the current frame.

3.4.3.2 Adaptive binary splitting protocol (ABS) [22] For fast identification in a frame, each staying tag retains its order of replying counter at the end of the last frame. Thus, the collisions caused by the staying tags can be totally avoided in the current frame. Each arriving tag changes its order counter to a random number from 0 to the total number of recognized tags of the last frame given by the reader. 61

CHAPTER 3

TAG ANTI-COLLISION PROTOCOLS

Thus, ABS can quickly recognize all of the arriving tags by equally distributing the arriving tags to different cycles. ABS can avoid collisions among staying tags, but it cannot prevent arriving tags from colliding with staying tags.

3.4.3.3 Blocking RFID anti-collision protocol for quick tag identification [44] It makes use of tag’s counters to save the order of reply for the recognized tags in the current interrogation session (frame), hence the tag preserve the identification order obtained from the last frame in order to avoid unnecessary collisions and idle cycles generated from identifying the staying tags in the current frame. It also avoids the collisions between the staying tags and the newly arriving tags. The reader starts the successive session by informing the tags by the number of recognized tags in the last frame. Hence, all arriving tags change their order counters to a random number larger than the number of the recognized tags in the last frame and smaller than the total predicted number of tags in the current frame by the reader estimation of the arriving tags.

3.5 Concluding Remarks In this chapter, we demonstrated the recent tag anti-collision protocols in literatures. However, the next chapter presents a survey for anti-collision protocols in multi-reader environment. Then, the proposed protocols will be stated and the comparison against the previous related work will show the performance. Now, Table 3.4 summarizes the advantages and disadvantages of the tag anti-collision protocols presented in this chapter.

62

CHAPTER 3

TAG ANTI-COLLISION PROTOCOLS

Table 3.4 Comparison among the tag anti-collision protocols.

Protocol Probabilistic Aloha based protocols; [11] (random access protocol) When collisions occurred, the tags that are involved in collisions wait a random time with some probability before re-transmitting their ID again.

Tree Slotted Aloha (TSA);[27] The behavior of the protocol follows a tree structure. It solves the collision as soon as it happens.

Basic Query Tree Scheme; [6] The tags whose prefixes match with the bits sent by the reader, replies back with their tag ID. If there is a collision, the reader queries for one bit longer prefix until no collision occurs.

Query Tree with Reversed IDs (QTR): Reversing the IDs of the tags and then applying the query tree (QT) protocol.

Query Tree-Based Reservation (RN16QTA)[14]: It starts by generating short16-bit random temporary IDS (RN16s) by all tags for giving the uniqueness to themselves, then applying the Query Tree Algorithm(QTA) with the temporary ID, first to be identified through its virtual ID, then to send its real ID.

Intelligent Query Tree Protocol (IQT)[6]: If all the items have first prefix bits common, then communication can take place using only the remaining (rem) bits.

Advantages

Disadvantages

*More efficient and faster than binary tree protocols when the number of tags is not very large and short ID length. * Memoryless.

*it cannot perfectly prevent tag collisions. * “tag starvation”: a probability of failing to read all tags in a limited time.

*It takes benefits of both Aloha and binary tree protocols. * Faster than Framed Slotted Aloha and query tree based protocols.

* Memoryless. * Requiring very less tag circuitry.

* TSA is not memoryless. * More complexity.

* Very slow.

* QTR protocol outperforms QT * Slow. protocol if the tags IDs are consecutive integers. * If the bit string has consecutive or identical prefix, it is effective to classify the suffix first. * It minimizes the problem of identification of full length ID to *Time consuming in transmitting the a problem of identifying temporary IDs first. colliding tags of randomly generated 16-bit IDS. * It reduces the overhead by using the temporary ID.

* Reduction in number of bits transmitted by tag. *Reduction in number of collisions by maintaining the history of tag read patterns. *Reduction in

63

CHAPTER 3

TAG ANTI-COLLISION PROTOCOLS

IQT exploits the existence of the common prefix patterns in the tag IDs

number of collisions for subsequent read cycles by using the information of previous read cycles* minimal change in tag hardware.

The collision tracking tree algorithm (CTTA) [25]:

RFID systems do not have to use the whole length of tag IDs for resolving each tag in the field of a reader.

CTTA is based on Query Tree Algorithm (QTA) except that this scheme uses collision tracking.

Binary Tree Searching and Balance Incomplete Block Design BIBD (4, 2, 1) [16]: The tag is divided into several sections, and each section (two bits) contains special sequences. Readers can get every ID by identifying each section one by one. * The same idea of Dived and Conquer; [26].

Dynamic Bit Arbitration (DBA) [17]:

*The searching speed of the algorithm is six times that of the Query Tree Protocol. * It is especially suitable for the identification conditions which contain numerous tags or tags with long ID.

* Simple and fast

If the collision occurs, the reader will record the position of the collision (k) in its register. The reader will choose a path by replying with 0 or 1.

ID-Binary Tree Stack Anti-collision Algorithm [19]: Only the tags whose stack counter are zero can take part in the later communication with the reader.

Each tag only transmits every bit of the ID once in the whole collision arbitration process. The tags communication complexity is equal to the length of the ID.

An Efficient Tree-Based * Minimize the total time Tag Anti-Collision Height-Oriented Protocol complexity. [20]: A height-oriented method of using an encoded binary number corresponding to the height of the vertex at the tag collision most recently occurred, which leads to prevent the redundant increase of the transmission bit length.

Adaptive Binary Splitting *ABS is an improvement of the 64

Due to multi-section processing, the identification time of this method is relatively high.

*overhead information, after identifying each tag, to inform the tags in the nearest collision node to change its current Quiet state to active state. The tags in active state need to know the last collision position to restart from.

*The need to send the height of the most recently occurred collision bit position after each identification to wake up inactive. * Reader command frame is relatively long per each transmitted tag bit. *Long delay.

CHAPTER 3

(ABS) Tag Anti-Collision protocol [21], [22]:

TAG ANTI-COLLISION PROTOCOLS

traditional binary tree protocol. It can avoid collisions among staying tags.

*It cannot prevent arriving tags from colliding with staying tags.

*It is good to restart from the most recent collision node; it is better than restarting in the tree root. * transmitting a part of ID starting from collision bit to the end of the ID, so decreasing the length of the time slot.

*Long reader feedback message to inform tags with the result of last timeslot (OBCT-

According to the reader’s feedback, all tags change its counter.

Enhanced Anti-collision Algorithm Based on Counter and Stack (EAA); [23]: The main assumption is the reader ability of truncating unnecessary data bits to reduce the receiving time at collision condition.

Bi-Slotted Tree based AntiCollision Protocols for Fast Tag Identification; [24], [25]: RFID reader transmits one ‘n − 1 length inquiring bits’ at a node which is related to a collided nth bit in the received tag IDs instead of two ‘n length inquiring bits’

Novel Anti-Collision Algorithm for Identifying Passive Tags (NEAA); [43]: It classifies the tags into sets according to the number of ones “1” in each tag-ID. Then it can identify tags set by set.

Blocking RFID anticollision protocol for quick tag identification; [44]: The reader starts the successive session by informing the tags by the number of recognized tags in the last frame. The tag preserve the identification order obtained from the last frame.

Identified-collision).

*In collision, it must feedback the position of the first collided bit.

*Very fast protocols. * BSQTA requires less average required bits for one-tag identification than QTA. * BSCTTA is the fastest identification protocol.

*It reduces the length of the time slot by truncating unnecessary data bits to minimize the receiving time. The reader does not need to receive any data after receiving the first collided bit.

* It has four type time slots with three bits instruction code, besides sending the position of the collided bit along with the collision code.

* It avoids unnecessary collisions and idle cycles * It avoids the collisions between the staying tags and the newly arriving tags.

* Similar to drawbacks of ABS. *long delay.

65

CHAPTER 4

ANTI-COLLISION PROTOCOLS FOR MULTI-READER ENVIRONMENT Abstract The purpose of this chapter is to provide a survey of Reader anticollision protocols. It presents the main classifications of anticollision protocols for multi-reader environment. The advantages and disadvantages of each method will be discussed.

CHAPTER 4

ANTI-COLLISION PROTOCOLS FOR MULTI-READER ENVIRONMENT

CHAPTER 4 ANTI-COLLISION PROTOCOLS FOR MULTIREADER ENVIRONMENT 4.1 Introduction In this chapter, we address the problem of collision in dense reader environment and its current solutions.

RFID systems are increasingly

being used in applications, such as those experienced in supply chain management, which require RFID readers to operate in close proximity to one another. Since the signal from a passive tag to the reader is a reflected signal, the read range of a reader is very limited. Hence, there is a need for integrating multi-reader in the building of RFID system for ensuring the coverage the whole existing tags in the environment, or the existence of mobile RFID where readers can be installed in a cellular phone.

Fig.4.1 Sources of reader collision problem

Working of overlapped multi-reader simultaneously over the same shared wireless channel leads to intersected interrogation zones. Readers physically located near one another may interfere with one another’s operation. Interference between readers leads to slow identification. Thus, it is a key issue to develop an efficient anti-collision protocol reducing collisions so as to identify all tags in the interrogation zone. Such reader collisions must be minimized to ensure the correct operation of the RFID

66

CHAPTER 4

ANTI-COLLISION PROTOCOLS FOR MULTI-READER ENVIRONMENT

system. Although both collisions modes (tags and readers collisions) increase the identification time, all previous researches deal tag and reader collision problems separately. Figure 4.1 shows the sources of reader collision problem. Introduction about the reader collision problem and its types are discussed in chapter 2. The most important factor in designing a protocol to avoid the reader collision problem was that the tags were passive and hence could not participate in collision avoidance. Secondly, adding any new functionality to the tags would increase the cost of the tags which would hamper large scale deployment of RFID systems. Hence we had to design a protocol that would not bring the tags in the picture; [32]. This chapter presents a survey of anti-collision protocols for multi-reader and multi-tag environment.

4.2 Main classification of reader anti-collision protocols Current solutions to the reader collision problems can be classified into two categories [29]: (1) Coverage based approach: the reading ranges of readers are adapted dynamically to reduce the overlapped area between adjacent readers as much as possible. This kind of approach increases the space re-used ratio, but it usually needs a central node to calculate the distance between every two readers and adjust their reading ranges, which will increase the complexity of realization and cost of the system. (2) Scheduling based approach: the available system resources, such as frequencies and time are allocated properly among readers to prevent readers from transmitting simultaneously. This kind of approach can reduce the possibility of reader collisions effectively, however it requires the system to establish and maintain information over the network, which will be time and energy consuming. Because of the low functionality of the passive tag, the tag-reader interference can be avoided only by having the 67

CHAPTER 4

ANTI-COLLISION PROTOCOLS FOR MULTI-READER ENVIRONMENT

neighboring readers operate at different times. Passive tags cannot distinguish between the communication signals from different readers. Therefore, when two or more readers may potentially communicate with the same tag, the readers must communicate at different times to ensure proper communication with the tag by each of the readers; [28]. All variants of the Frequency Assignment Problem assume mobile devices, e.g., cellular telephones, are intelligent and capable of aiding in the communication process and, in particular, are capable of distinguishing between and selectively communicating on different frequencies. An RF tag, however, is not capable of aiding in the communication process. It is not able to distinguish between different frequencies; therefore, a tag responds to all communications that it is able to detect. This lack of functionality on the part of the mobile device, i.e., the tag, gives rise to the tag interference experienced in the Reader Collision Problem. Unlike some cellular base stations, RFID readers are only capable of communicating on a single frequency at a time. Furthermore, since RFID systems operate in the ISM (Industrial, Scientific, Medical) frequency bands, the frequency used for communication cannot be controlled, only the time at which the readers communicate may be controlled. This is in contrast to cellular networks operating in privately owned frequency bands that permit explicit control of frequency assignments. The potentially high required communication rate with tags in its vicinity requires interference caused by and experienced by readers to be minimized.

Hence, most of the proposed protocols go around TDMA scheduling to prevent readers from transmitting signal simultaneously.

68

CHAPTER 4

ANTI-COLLISION PROTOCOLS FOR MULTI-READER ENVIRONMENT

4.3 Standard multiple access mechanisms [32] It cannot be directly applied to RFID systems due to the following reasons: - FDMA: With FDMA, the interfering readers use different frequencies to communicate with the tags. Since the RFID tags do not have any frequency selectivity, they

cannot select a particular reader frequency for

communication. Hence FDMA is not a practical solution in RFID systems. - TDMA: With TDMA, the interfering readers are allotted different time slots thus avoiding simultaneous transmissions. However this is similar to the well known coloring problem in graph theory. And because of mobility, non interfering readers may move closer and start interfering. Hence a fixed TDMA based protocol may not be very efficient with mobility. - CSMA: the read ranges of two readers may not overlap. However, the signals from the two readers can interfere at tag (hidden terminal problem). This case can also happen when the two readers are not in each other's sensing range making carrier sensing ineffective in RFID networks. - CDAM: CDMA will require extra circuitry at the tag which will increase the cost of the tags. Also code assignment to all the tags at the deployment site may be a complicated job. Hence CDMA may not be a cost effective solution.

4.4 European Telecommunications Standards Institute (ETSI) EN 302 208 (Listen-Before-Talk (LBT)) [31] It has a CSMA based protocol called "Listen Before Talk". The reader first listens on the data channel for any on-going communication for a specified minimum time. If the channel is idle for that time, it starts reading the tags. If the channel is not idle, it chooses a random backoff. But the readers may not be able to detect collision by carrier sensing alone.

69

CHAPTER 4

ANTI-COLLISION PROTOCOLS FOR MULTI-READER ENVIRONMENT

4.5 Colorwave: An Anti-collision Algorithm for the Reader Collision Problem [28] It is a distributed TDMA based, where the time is divided into frames and a frame is divided into a number of slots. Each timeslot is divided into a long reader-tag communication period and a short reader-reader kick period. Each reader selects any time slot (color) randomly and reads its tag in its own time slot. If there is a collision, the reader selects a new time slot and sends a kick (small control packet) to all its neighbors. The frame size can be increased or decreased according to percentage of collisions. If the current color of a reader is the same as the time slot, the reader checks whether a collision with neighboring readers occurs or not. When there is no collision, the reader can conduct tag identification during the time slot. And it reserves the same color in the next frame. In the case of a collision, the collided readers do not identify tags in the current frame, and they randomly reselect and reserve their colors for the next frame. After the reselection, they broadcast the changed colors to interfering neighbor readers in order to prevent them colliding in the next frame. This reselection and reservation action is referred to as a kick. The readers receiving the kick information compare their colors for the next frame with the kick information. If they are the same, they randomly change the colors to different colors among the max color; [30]. However, if the number of the adjacent readers is much larger than the max color, many readers may experience collisions. On the contrary, when the max color size is greater than the number of readers, many colors may be wasted. Thus, to improve the throughput of RFID systems, it is necessary to adjust the max color appropriately depending on the number of neighbors. The distributed nature of the algorithm allows each reader to minimize collisions based

70

CHAPTER 4

ANTI-COLLISION PROTOCOLS FOR MULTI-READER ENVIRONMENT

upon local information. Global communication and information sharing are not required. If the percentage of successful transmission goes below certain threshold, the maxColors is incremented and if the percentage increases beyond certain threshold, the maxColors is decremented. However Colorwave requires time synchronization between readers. Also, Colorwave assumes that the readers are able to detect collisions in the RFID system. However it may not be practical for a reader alone to detect the collisions that happen at the tags unless the tags take part in the collision detection; [32].

4.6 Channel Monitoring Algorithm [33] Colorwave can reduce reader collisions by adjusting the frame size according to the collision probability. However, a reader randomly chooses the time slot in the next frame, which interrupts the tag reading of another reader choosing the same time slot, and a reader collision occurs. In the channel monitoring algorithm, each reader monitors the slots in the frame and chooses the slot with the minimum occupied probability, which is defined as the percentage of readers that choose a certain time slot among all readers. Thus, collisions caused by randomly choosing a time slot are reduced in the channel monitoring algorithm.

4.7 Solution to the Reader Collision Problem based on Central Cooperator (CC)-RFID [29] In RFID systems, the tags (lying in overlapped reading range) read by adjacent readers contain exactly the same information. However, none of the current reader collision protocols takes this information redundancy

71

CHAPTER 4

ANTI-COLLISION PROTOCOLS FOR MULTI-READER ENVIRONMENT

into consideration. The tag will be read several times by each reader in its range.

Fig.4.2 CC-RFID System Architecture

If the same tags could be read by one reader and shared with others, the reading efficiency will be further increased. Based on this fact, Central Cooperator (CC)-RFID (Figure 4.2) is used to improve the performance of multiple readers RFID system. It converts the present ‘multiple points to multiple points’ (MP2MP) collision problem into two ‘multiple points to one point’ (MP2P) classical collision problems. The reading queries from several readers could be multiplexed by CC and the same tag information could be stored and shared among adjacent readers. Additionally, CC-RFID does not change the air interface and can be easily applied in current RFID system. Simulation results show that the reading speed and efficiency of CC-RFID scheme is highly improved over current popular PULSE protocol. The main shortcoming of CC-FRID is the requirement of a special device called Central Cooperator.

72

CHAPTER 4

ANTI-COLLISION PROTOCOLS FOR MULTI-READER ENVIRONMENT

4.8 RFID Reader Anti-collision Algorithm Using a Server and Mobile Readers Based on Conflict-Free Multiple Access (CFMA) [30]

Fig.4.3 An example of a reader network model in CFMA

It uses a central server which receives reader request with its location information. The server decides whether a reader is safe to operate without collisions with currently operating neighbors or not as shown in Figure 4.3. But, it assumes that every reader can be aware of its location accurately. It uses two channels, a channel for tag identification and a channel for communications between a server and readers. In the proposed CFMA, a RFID server grants a service to readers for tag identification on a first come-first-served basis. If the requests of some readers arrive at the server simultaneously, the server randomly arranges the order of service for the readers. When a reader desires tag identification, the reader transmits a request command with its location information to the server, and it looks forward to receiving an answer of the server. On receiving the request command of a reader, the RFID server compares the reader’s location with the readers of currently operating, i.e., identifying tags. Then the server decides whether the requesting reader is able to operate without interfering with the other operating readers or not. If there is no neighboring readers 73

CHAPTER 4

ANTI-COLLISION PROTOCOLS FOR MULTI-READER ENVIRONMENT

interfering with the requesting reader among the operating readers, the server sends an accept command to the reader to permit communications with tags. If the server concludes that several readers may be interfered by the operation of the reader, it does not respond to the request of the reader. And the reader waits for the accept command until the max waiting time. CFMA avoids reader collisions effectively because a server notifies readers whether they can operate or not from the location information of readers. It defines the efficiency as the number of simultaneously operating readers per millisecond in an area.

4.9 Pulse Protocol [32] The communication channel is divided into two parts: data and control channel. A reader communicates with tags via the data channel (reader-tag communication),

and

uses

the

control

channel

(reader-reader

communication) to send a beacon signal, which notifies other readers to avoid collisions as shown in Figure 4.4. The reason for a separate control channel is that, a beacon transmission at a higher power on the data channel will interfere with any ongoing reader-tag communication in the interference region of this beacon transmission. Data channel is the main communication channel, accounting for the bulk of the spectrums. While control channel has little communication of data, it occupies fewer spectrums. The size of the scope of control channel must ensure all the readers that may collide with each other can communicate by control channel by raising the transmitting power of control channel. If a reader receives a beacon, it waits for a certain time, and then rechecks the control channel, until no beacon is received from other readers.

The

communication range in the control channel is such that, any two readers

74

CHAPTER 4

ANTI-COLLISION PROTOCOLS FOR MULTI-READER ENVIRONMENT

that can interfere with each other on the data channel (channel used to read the tags), are able to communicate on the control channel.

Fig.4.4 Example of Pulse protocol (note: R2 active, R1 silent)

„ Before communicating, a reader listens on the control channel for any beacon for Tmin time. „ If no beacon on the control channel for Tmin , start communication on the data channel. „ Reader periodically transmits a beacon on the control channel while communicating with the tags. „ As shown in Figure 4.5, after Tmin time has elapsed and it did not receive any beacon for that time, the reader concludes that there is no other reader in the neighborhood which is reading the tags. Hence it enters a contention phase and chooses a random backoff time (contend backoff ) from the interval [0. . .CW]. If it chooses i, it waits for i beacon intervals in state CONTEND. If it now receives a beacon, it has lost this cycle and waits for the next cycle, i.e. until it does not receive a beacon for at least Tmin time. If the randomized backoff time is over and the reader did not receive any beacon, the reader assumes that there is no other reader to compete and hence it sends a beacon on the control channel and starts communicating with 75

CHAPTER 4

ANTI-COLLISION PROTOCOLS FOR MULTI-READER ENVIRONMENT

the tags on the data channel. This randomized backoff helps to avoid collisions between readers; otherwise many readers would try to transmit the beacon simultaneously after waiting for Tmin time. Contend backoff is a multiple of beacon intervals to improve fairness.

Fig.4.5 Example of randomized backoff times in pulse protocol

Fig.4.6 PULSE Protocol Flowchart

Figure 4.6 shows the flowchart of the Pulse protocol. While the reader is communicating with the tags, the reader sends a beacon on the control channel every beacon interval. This beacon acts as a notification to the 76

CHAPTER 4

ANTI-COLLISION PROTOCOLS FOR MULTI-READER ENVIRONMENT

neighboring readers so that they can withhold their communication with the tags and thus avoid possible collisions. Raising the transmitting power of control channel can broader its scope, which can ensure the scope of control channel is larger than data channel in order to avoid all types of reader collision including tag hidden terminal problem. After the communication with the tags is over, the reader again waits in the WAITING state and the cycle continues. It is a distributed protocol for an RFID network which uses a beaconing mechanism by sending periodic beacon on the control channel. The protocol is simple. It mitigates the reader collision problem. It reduces the reader collisions and also increases the read rate of the system. It requires very less overhead on the reader side and absolutely no support on the tag side. It is also very effective in a mobile scenario facilitating the use of mobile readers.

4.10 Enhanced Pulse Protocol using Slot Occupied Probability in Dense Reader Environment [34] In the conventional pulse protocol algorithm, when a reader generates a new random backoff delay, there is a probability that the backoff delay time is the same as that of another reader. Beacon can collide with another beacon form another reader in control channel. Thus, a reader collision occurs as shown in Figure 4.7.

Fig.4.7 Reader collision in the conventional pulse protocol

77

CHAPTER 4

ANTI-COLLISION PROTOCOLS FOR MULTI-READER ENVIRONMENT

The improved pulse protocol reduces that type of reader collisions as shown in Figure 4.8. It improves the conventional pulse protocol by introducing slot occupied probability. The implementation of this improvement is simple, yet it effectively mitigates most reader collisions in dense reader mode. Slot Occupied Probability (SOP) is defined as the percentage of readers that choose the same time slot all active readers, as given in equation 4.1.

In the conventional pulse protocol algorithm, it is assumed that power of the beacon signal is boost-up enough to be received by all of neighboring readers. By communicating with other readers via the control channel with beacon signal, which includes back-off time information, a reader can easily find the backoff delay time of the other readers.

Fig.4.8 Mitigating reader collision by Slot Occupied Probability

Thus, whenever it generates a random backoff delay, the reader calculates the SOP for the backoff time slot and compares the value with zero. If SOP is larger than zero, this means other readers are supposed to communicate with tags during the same time slot. Thus, the reader generates another backoff delay for avoiding collisions with other readers during that time slot.

Before communicating with the tags, a reader remains in the waiting 78

CHAPTER 4

ANTI-COLLISION PROTOCOLS FOR MULTI-READER ENVIRONMENT

state for a minimum time Tmin, to ensure no beacons are received during this time. If no beacon signal is received by the reader during Tmin, it considers that no other neighboring readers are communicating with tags. Thus, it enters the contend state. Subsequently, the reader chooses a random backoff time (contend_backoff), e.g., n time slots, between zero and the maximum backoff delay, and calculates the value of SOP for the nth time slot. If SOP is equal to zero, the reader does not need to change the random backoff time. If SOP is larger than zero, this means there are other readers that are supposed to communicate with tags during that time slot. Hence, the reader changes its backoff time to another random value, to avoid collisions with other readers, until SOP is equal to zero. If a beacon signal is received by the reader during the backoff time, it returns to the waiting state until no beacon is received for at least Tmin time. If the random backoff time has expired and no beacon is received by the reader, it sends a beacon via the control channel and starts communicating with the tags via the data channel. The beacon notifies the neighboring readers that they should delay communication with the tags and thus avoid possible collisions. The improved pulse protocol reduces the identification time compared with the conventional channel monitoring algorithm and pulse protocol, when the number of readers is larger than one. This verifies that the algorithm reduces the probability of collisions. The improvement in the identification time by the Enhanced Pulse protocol is proportional to the number of readers compared with the conventional pulse protocol. The simulation results show that the algorithm reduces the identification time, and increases the system throughput compared with the conventional reader anti-collision algorithms as shown in Figures 4.9 and 4.10. 79

CHAPTER 4

ANTI-COLLISION PROTOCOLS FOR MULTI-READER ENVIRONMENT

Fig.4.9 Identification time of the Enhanced Pulse protocol and conventional algorithms

Fig.4.10 System Throughput of the Enhanced Pulse protocol and conventional algorithms

4.11 Pulse Protocol with reduced energy consumption [35] In Pulse protocol, the active reader will send Beacon package in every interval during the whole process of the communication between tags and the reader, until the process is finished. That protocol does not need the working reader to send the Beacon package periodically. So that readers can further reduce energy consumption.

80

CHAPTER 4

ANTI-COLLISION PROTOCOLS FOR MULTI-READER ENVIRONMENT

The main protocol assumptions: • Channel is divided into two parts: control channel and data channel. Data Channel is for communications between tags and readers and control channel is for communications between readers and readers; there is no interference between data channel and control channel. • Readers can receive the data from data channel and control channel at the same time, and it can also send the data to control channel and data channel. • Readers’ movement is unrestricted, which means readers can in-out the whole network freely. • There are three reader signals that can be transmitted through the control channel: REQ, BUSY, and END signal. The reader does not send the data packages immediately but intercept whether there are REQ or BUSY packages from others. If there are REQ or BUSY packages from others, the reader enters the WAITING state. The time in WAITING state should not be less than the time in which another reader can finish the communication with tags, unless the reader receives the END packages from others. On the contrary, if there are no REQ or BUSY packages from others, the reader will send its REQ package in control channel. REQ is a broadcast signal and all the neighbors can receive it. If there are some neighbors who are communicating with tags, they will send BUSY packages to the reader when they receive REQ package from the reader. If the reader receives BUSY packages from others, it will enter the WAITING state too. Otherwise, the reader will begin to communicate with tags which are in the interrogation zone of it. When the process of communication is finished, it will send END package

81

CHAPTER 4

ANTI-COLLISION PROTOCOLS FOR MULTI-READER ENVIRONMENT

to control channel. END signal is the same to REQ and BUSY signal. They are all broadcast signals which can ensure neighbors to receive them. That protocol achieves performance similar to that of the Pulse protocol, but it does not send the Beacon package periodically. So that readers can further reduce energy consumption, which is suitable for mobile readers.

4.12 GENTLE: Reducing Reader Collision in Mobile RFID Networks [36] The basic idea is that readers avoid multiple reader-to-tag collision by using beacon messages when they are close one another and reader-toreader collision by using multi-channel when the distance between those readers is long. The way that a beacon message is sent to other neighbor readers is basically the same as in Pulse protocol but the main difference of GENTLE protocol to the PULSE protocol is that: •

In PULSE, a reader periodically broadcasts a beacon message to much

further distance (i.e., further than the scope of interference range on a separate control channel, about several hundred meters) than Gentle protocol in order to avoid all types of reader collision including tag hidden terminal problem.

Fig. 4.11 Example of PULSE operation 82

CHAPTER 4

ANTI-COLLISION PROTOCOLS FOR MULTI-READER ENVIRONMENT

As shown in Figure 4.11, a reader R1 sends a beacon message periodically far away at control channel, and many neighbor readers including R2 listen to the message at the shared control channel. The reason why the reader R1 sends the beacon message up to R2, which is far away from the R1, is because R2 may affect the tag T1 communicating with R1 (i.e., tag hidden terminal problem). To avoid the interference, accordingly, the readers R2, R3, R4, R5, and R6 have to wait until they do not receive beacon message from R1 any more. The reader R1 reads tag T1 and T2 at data channel while neighbor readers wait. In this way, long waiting time for reading tags is required, which will result in reduction of read throughput through the whole network. In [36], there is a verification that tag hidden terminal problem can not exist in real world with the RFID path loss model. Hence, there is no need for sending a beacon messages further than the scope of interference range of readers to avoid hidden terminal problem.

Fig. 4.12 Example of GENLE protocol operation

83

CHAPTER 4

ANTI-COLLISION PROTOCOLS FOR MULTI-READER ENVIRONMENT

• In GENTLE, It reserves one channel as a control channel that all readers can share. The active reader communicating with a tag, periodically sends a beacon message through the control channel. But, the beacon message is restricted to be sent only within a certain scope, i.e., only up to the distance that multiple reader to- tag collision might occur. As shown in Figure 4.12, the reader R1 broadcasts a beacon message only up to R3 and R4, which might cause multiple reader-to-tag collision to the tag T1 and T2. The tag T1 can identify R1's signal even in case that both R1 and R2 transmit signals simultaneously since R1's strong signal is superior to very weak one of R2. Thus, only R3 and R4 wait while R1 is reading T1 and T2 at channel 2. On the other hand, R5 and R6, which are located in R1's interference range but have not received the beacon message, randomly choose empty data channels (ch. 3 and 4) to avoid reader-to-reader collision and read their own tag, T3 and T4. Since R2 is outside of R1's interference range, there is no reader to-reader collision problem even though R2 uses the channel 2 of R1. In this way, multiple reader-to-tag collision problem can be solved with sending beacon messages when a reader is closely located in with an interfering reader and when the distance between those readers is long, the readers might use multi-channel that leads to improvement on read throughput. Further, PULSE can not benefit from multi-channel whereas the Gentle protocol helps increase read throughput using multi-data channel. • Forwarding mechanism (Gentle 2) The reader reading tags broadcasts the data of a tag by a beacon message to neighbor readers whenever it reads each tag. This mechanism makes readers save waiting time for reading tags and decrease the number of attempt to read tags and reduce reader collision that can occur in future. 84

CHAPTER 4

ANTI-COLLISION PROTOCOLS FOR MULTI-READER ENVIRONMENT

It increases read throughput of all readers when there is high probability of multiple reader-to-tag collision and the data of tags do not require particular security.

4.13 DiCA: Distributed Tag Access with Collision-Avoidance Among Mobile RFID Readers [37] It is similar to pulse protocol. For the sake of saving energy they don’t periodically broadcast message like Pulse. Of course, the energy is very important factor for mobile RFID reader. It implies a distributed and energy efficient collision avoidance algorithm. As in pulse, DiCa also has data channel and control channel. Each reader contends through control channel and the contention winner reads tags through data channel and other wait until the channel is idle. However, DiCa has some shortcoming. It requires sufficient time to exchange the contention message. It tries to prevent the collision after it takes place rather than acting actively at the first sight. So, it does not reduce the collision problem efficiently.

4.14 Multi-Channel MAC Protocol (MCMAC) [38] MCMAC works in a similar manner to the conventional LBT (Listen Before Talk). It adds N channels (Multi-Channel) in each RFID reader, and all channels have the same bandwidth, none of channels overlap, so the packets transmit on different channels do not interfere with each other. The total N channels are composed of N-1 data channels and one control channel. The control channel can be used to exchange the control packets and data packets. But the data channels can only deliver the data packets.

85

CHAPTER 4

ANTI-COLLISION PROTOCOLS FOR MULTI-READER ENVIRONMENT

The communication range in the control channel is such that, any two readers that can interfere with each other on the any data channel (channel used to read the tags), are able to communicate on the control channel. The protocol is present only at the reader since the tags do not take part in the collision avoidance. MCMAC broadcasts control message after it wins contention in a control channel and gains access to the data channel. The control message informs other neighboring readers within the interrogation zone that the particular channel is occupied for a certain time. After receiving a control message from neighboring reader, the other readers do not use that channel for a certain period of time and try to gain access to another channel. Despite the fact that this approach can mitigate the reader-to-reader problem, it cannot solve the reader-to-tag problem. Passive RFID tags are unable to discriminate between tow data channels. Therefore, multiple data channels are basically not applicable in a passive tag environment. [41]

4.15 A Reader Anti-collision MAC Protocol for Dense Reader RFID System (RAMP) [39] (Multiple Data Channels and Estimated Channel Hopping) In MCMAC[38] hopping is performed in random fashion, which is not efficient way to search idle channel because reader might perform continuous channel hopping for a long time in order to find idle channel when channel utilization is high. That protocol extends the pulse protocol by adding multiple data channels and the channel hopping algorithm under estimated rule for hopping. The channel hopping algorithm helps to decide whether to hop for new channel or wait in the same channel. The channel hopping decision is made on the basis of density of the readers. The density

86

CHAPTER 4

ANTI-COLLISION PROTOCOLS FOR MULTI-READER ENVIRONMENT

of the readers is an estimable way to decide either to hop for new channel or just wait in the same channel. If the number of data channels is less than the number of readers then hopping is better than waiting in the same channel. It achieves efficient channel utilization.

4.16 An Efficient MAC Protocol for Throughput Enhancement in Dense RFID System (Anti-Collision MAC (ACMAC)) [40] It assumes that there are N = {c1, c2, ..., cn} number of data channels and a control channel. Data channels are for the readers-to-tag communication and control channel is for the reader-to-reader communication. Readers can transmit control signal through the control channel while communicating with tags through the data channel. Each reader listens before talk as in conventional LBT. The reader wants to start communication enters into the Listening Stage (LS) for TL time to win contention in a particular data channel. In the LS, * If the reader receives any beacon message (in the control channel) from any of neighboring readers, it determines that the data channel is not idle. Now, it selects a backoff counter within contention window (CW) and waits until it becomes zero. * If reader in LS does not receive any beacon message within TL time, it enters in the Waiting Stage (WS) and waits for Tw time. In the WS, *if reader receives beacon message, the reader has to decide whether to wait in the same channel until that channel gets free or hop to another idle channel based on probability based channel selection algorithm. *If the reader does not receive any beacon message in WS, the reader broadcasts the beacon message to the neighbors in the control channel and 87

CHAPTER 4

ANTI-COLLISION PROTOCOLS FOR MULTI-READER ENVIRONMENT

occupies the data channel. After broadcasting beacon message, reader waits for Tw time and starts communication with tags in the data channel. To avoid the possible collisions due to accessing the control channel by multiple readers at the same time, it uses random backoff mechanism. Each node that wants to communicate should contend the control channel with random backoff. They select a backoff counter within a contention window (CW). CW starts from the minimum value (Bmin) initially and the counter is decremented. *If the control packets collide in control channel with the packets from neighbor nodes, both of the colliding nodes increase their CW size multiplying by two. The reader that with the lowest slot size can occupy the channel and other nodes freeze the counter until next contention period. A good algorithm is able to reduce the probability of collision so that it consumes less time to identify tags and achieves higher system throughput(number of tags read per time unit) as shown in Equation 4.2. Figure 4.13 shows the network throughput comparison between the current published protocols.

Fig.4.13 Network throughput.

88

CHAPTER 4

ANTI-COLLISION PROTOCOLS FOR MULTI-READER ENVIRONMENT

Figure 4.14 shows the flowchart of working principle of ACMAC.

Figure 4.14 Flowchart of working principle of ACMAC.

• At that point, we have presenting wide range of anti-collision protocols for both tag and reader collision problem. • In the next chapters, we will introduce our proposed protocols for solving the collision problem in multi-tag multi-reader environment with the comparative results to highlight the contributions of this thesis. 89

CHAPTER 5

PARALLEL BINARY SPLITTING (PBS) PROTOCOLS

Abstract This chapter introduces the proposed tag anti-collision solution based on the Parallel Binary Splitting (PBS) approach. The advantages (simplicity & low overhead) of the new PBS searching path will be clarified. Then an enhanced Fast Parallel Binary Splitting (FPBS) is introduced. FPBS depends on full duplex tag ability. FPBS Performance, in successive reading rounds, will be studied. Moreover, the application of PBS on multi-reader environment is introduced in integrated reader and tag anticollision protocol based on Similar Topology Trees (STT) in one collision domain. SST protocol is designed to overcome all sources of collision. The basic key in PBS protocol and FPBS protocol is the simple dialog between the reader and tags.

CHAPTER 5

PARALLEL BINARY SPLITTING PROTOCOLS

CHAPTER 5 PARALLEL BINARY SPLITTING PROTOCOLS 5.1 Introduction In this chapter, we present a new tree based anti-collision algorithm for Radio Frequency Identification (RFID). We try to satisfy the research question which is: “How to setup anti-collision algorithm that needs lower bit rate transmission and simple logic reconstruction?”. The proposed algorithm is based on parallel binary splitting (PBS) technique which follows a new identification path through the binary tree.

The dialog

between the reader and tags needs only one bit tag response followed by one bit reader reply (one-to-one bit dialog). After identification of each tag at a leaf node of the binary tree, the new path does not need the restarting node position in the binary tree. The protocol relies on two counters and two registers in tag side for setting up self transmission control. In the collision state, tags do modify their next replying orders in the next bit level. The novelty in this protocol can be found in developing the parallel node identification path, the minimized reader response of one bit as comment to inform the tags with the collision state of each bit, and the minimum overhead needed for the identification. The system works with one- to- one bit dialog to build the tree nodes.

So, the number of

transmitted bits is equal to twice the number of the binary tree nodes of the existing tags except the leaves nodes. Moreover, this chapter describes Fast Parallel Binary Splitting (FPBS) as enhanced version of the PBS protocol to reduce the identification time to half of the PBS. In the end of this chapter, we present an integrated protocol for anti-collision for multitag multi-reader environment depending on the parallel splitting approach. It depends on building Similar Topology binary Trees (STT) for all readers.

90

CHAPTER 5

PARALLEL BINARY SPLITTING PROTOCOLS

5.2 Parallel Binary Splitting (PBS) Protocol for Tag Anti-collision The main objective of the proposed algorithm is to minimize the reader transmitted information to only one bit feedback, informing the tags with collision or no collision condition. It does not need query of last collision bit position. Hence, reduction in the number of transferred bits between the reader and the tags is achieved. The proposed method is based on the parallel binary splitting (PBS) that dynamically modify the tag relative reply order with respect to the adjacent tags. The proposed protocol utilizes self assigning relative virtual orders, which determines tags transmission order in a dynamically updated order queue. Only one bit reader response is used to minimize the amount of transmitted data by the reader, (1: means collision, 0: means no collision). The reader informs tags with the result of tags bit transmission, hence enabling the tags to continually modify its recognized order in the reply queue. Therefore, the tag can determine its future replying order and setting up a self transmission control. All existing tags are completely identified according to its relative order. The parallel splitting approach outperforms most of the recent depth first search (DFS) techniques [17], [19], [21], [22], [23] in most cases. Figure 5.1 shows the difference between the new parallel binary splitting (PBS) technique and the traditional depth first search (DFS) technique which enters each new splitting path until ending at one tree leaf. The main objective of the PBS is to simplify the dialog between the reader and tags during the reading process (the reader extracts the tags' IDs). In the new followed PBS path, the advance in the tree appears as parallel advance. The reader listens to tags' one bit at each node in the same bit level before the advance to the next discovered level as shown in Figure.5.1.b. PBS does not need query of last collision bit position to restart another path such as in DFS (Figure.5.1.a). The details of the proposed protocol are described next. 91

CHAPTER 5

PARALLEL BINARY SPLITTING PROTOCOLS

Fig.5.1 The difference between parallel splitting identification path and the depth first search path

5.2.1 Reader operation Because we can’t depend on tags to sense transmission from adjacent tags, the reader operates as repeater that broadcasts the result of successive tags transmission bit-by-bit, (Is it collision or no collision?). Reader and tags start the identification session by assuming that there is one tag. Each tag assumes that it is the only existing tag with all its counters and registers will reset. The Reader operation can be summarized as follows: 1- Continually updating the pretended binary tree of existing tags according to tags reply. 2- Exploring the future nodes from each already discovered node (path) of the previous splitting level, by examining (scanning) the previously discovered paths in the past binary splitting level as shown in Figure 5.1.b. The advance in the binary tree appears as parallel advance. 3- Receiving the marked bit of each subgroup in the predetermined order. Each tag replies to the reader interrogation in its previously replying order by sending its marked bit. 4- Detecting the state of the last tag reply, in the scanned subgroup, such that:

* Detecting Collision

: send “1” as collision report.

* Detecting No Collision : send “0” as collision free report. 92

CHAPTER 5

PARALLEL BINARY SPLITTING PROTOCOLS

Each tag knows its current replying-order, and remains in its order in the replying queue until detecting a tag collision. The tags will classify themselves in a new subgroup according to the reader signal only.

5.2.2 Tag operation PBS protocol is based on remapping the discovered binary tree configuration after each tag and reader one bit dialog (one bit reply). By knowing the state of the last tags reply from reader report, each tag continually changes its relative position in virtual reply queue. Each tag in the reply queue waits its reply order to send the next level ID-bit. Tags use simple logic operation based on two counters and two registers, to achieve self transmission control and dynamically updated replying orders.

The registers and counters are defined as follows: 1. Current Path Register (CPR): is used to store the current number of paths (binary branches) in that bit level. It contains the number of checked nodes in that binary level. 2. Next Paths Counter (NPC): is used to track the total number of continually discovered paths. It will be incremented when the reader reports a collision tag reply. Any collision means an increasing in the binary braches by one. CPR will be loaded by NPC content at the end of each bit level splitting. 3. Current Order Register (COR): is used to store the tag replying order with respect to the current number of paths in CPR. 4. Next Order Counter (NOC): is used to track the changing in the tag replying order. It will increment when a new branch of lower order appears in the binary tree. COR will be loaded by NOC content at the end of each bit level splitting. 93

CHAPTER 5

PARALLEL BINARY SPLITTING PROTOCOLS

The tag operation can be described as follows: 1. Receiving the reader starting command. 2. Initially, each tag starts by thinking that it is the only tag in the reader range and resets the two counters and the two registers. 3. One-bit tag response in its replying order, one-bit reader report will follow that. 4. Scanning the previously discovered paths (nodes) in the past splitting level. 5. Each subgroup sends the current marked bit in the current bit level. 6. Tags know the collision state from the reader report at each node. 7. Tags modify its control counters as follow: a- IF "No Collision": THEN, NO change in its order and the total number of paths. b- IF "Collision" :

THEN

*increment the total number of paths (increment NPC). * “IF {“the tag is not scanned in the current bit level (i.e. it is waiting its replying order)”

OR “it is the tag replying order and participating in the

current tag collision by sending its marked bit which is one”} THEN:

incrementing its replying order in the next splitting level

(increment NOC). 8. Registers (COR, CPR) are updated by the contents of the corresponding counters (NOC, NPC). Where the update in the paths and orders will not be considered until the start of the next splitting binary level. 9. Scanning the next splitting level and repeat the process starting from step 4 until completing "n" level of the ID length.

94

CHAPTER 5

PARALLEL BINARY SPLITTING PROTOCOLS

(PBS)Procedure Executing on the Tag ----------------//* initialization *////* each tag think that it is the only tag. Current Order Register: COR=1, Current Path Register: CPR=1, Next Order Counter: NOC=1, Next Paths Counter: NPC=1, Bit Marker (or pointer): M=1, MB: Marked Bit, n: n: length of tag ID bits=96 level R: Received feedback bit from reader //* , that inform the result of tag responses (R=0: no collision, R=1: collision result) -----------------------------------//* tag operation *//-

Waiting (tree start command) { //*repeating the loop (n) turn: number of tag ID bits (no. of levels).

For L1= 1: n //* looping to scan the current checked paths in that level

For L2=1: CPR //* is it the tag order to reply?

IF COR (current order register) =L2 //* yes, it is my order to reply.

{Send the marked bit} Else //*no, it is not tag order, it should stay without response

{Let others tags to reply one bit period without interfering them} End if //* receiving reader report for each tag one bit response

{Receiving result feedback of current checked node} //* is it a collision feedback reply?

IF

R= 1

Then

//* increment number of checked points in the next level

Increment NPC (Next Paths Counter) //*IF {the tag is not scanned in the current bit level (i.e. it is waiting its replying order)} OR {“It is the tag replying order and participating in the current tag collision by sending its marked bit which is one”}

IF {(COR > L2) or (COR = L2 & MB = 1)} Then //* incrementing its replying order in the next splitting level.

{Increment (NOC) Next Order Counter) ; End if End if; End for L2; //* Registers (COR, CPR) is updated by the contents of the corresponding counters (NOC, NPC). And Increment the marker to point to the next bit. For scanning the next splitting level

COR =NOC, CPR =NPC, increment M End for L1} Fig.5.2 PBS procedure executing on the Tag

95

CHAPTER 5

PARALLEL BINARY SPLITTING PROTOCOLS

Figure 5.2 shows a detailed description for the self transmission control in the tag side by two counters and two registers.

5.2.3 Performance analysis

Fig.5.3 The diagram of four tags binary tree with the exchanged bit stream

In this section, the performance of PBS algorithm will be discussed. To demonstrate the tag identification, Figure 5.3 shows the diagram of four tags in binary tree, as an example, {A, B, C, D} = {0000, 0110, 1110, 1111} to be identified and, the bit stream between reader and tags, with each node contains its order of reply according the PBS path. Reader reconstructs the binary tree by scanning the tree nodes in the shown order, and receiving tag responses at each node. Figure 5.3, Figure 5.4, and Table 5.1, describe in details, the process of node exploration and order modification.

96

CHAPTER 5

PARALLEL BINARY SPLITTING PROTOCOLS

Fig.5.4 Parallel splitting scan for path exploration

Table 5.1 The anti-collision process of the proposed PBS algorithm (--: silent) Tags Registers & Counters

Reply of TAG Reader step

(After Reader Response)

COR

NOC

Reply A (0000)

B (0110)

C (1110)

D (1111)

A

B

C

D

A

B

C

D

CPR

NPC

(for all tags)

(for all tags)

a

Start

0

0

1

1

1

1

1

1

1

1

1

1

1

1

b

1

0

1

----

----

1

1

2

2

1

1

2

2

2

2

c

1

----

----

1

1

1

1

2

2

1

2

3

3

2

3

d

0

0

----

----

----

1

2

3

3

1

2

3

3

3

3

e

0

----

1

----

----

1

2

3

3

1

2

3

3

3

3

f

0

----

----

1

1

1

2

3

3

1

2

3

3

3

3

g

0

0

----

----

----

1

2

3

3

1

2

3

3

3

3

h

0

----

0

----

----

1

2

3

3

1

2

3

3

3

3

i

0

----

----

0

1

1

2

3

3

1

2

3

3

3

3

j

1

----

----

----

----

1

2

3

4

1

2

3

4

4

4

97

CHAPTER 5

PARALLEL BINARY SPLITTING PROTOCOLS

The number of tree nodes (except the leaves node) = 9 nodes. It consumes one bit for each node for tag response and one bit for reader to report the type of each node (collision or no collision). * Number of bits transmitted by tag = Number of bits transmitted by reader = 9 bit. * The overall bit transferred between tags and reader= 9 bit tag response + 9 bit reader reply

= 18 bit.

The performance of the proposed mechanism is compared with the EAA algorithm in [23]. In EAA, it uses 19 feedback + 11 tag responses = 30 bits to identify the same number of tree nodes in our example. Moreover, in the proposed algorithm, the relative order of each tag will be saved; and less number of transmitted bits to make another identification session will be achieved, (it will be discussed in section 5.3.5).

Figure 5.5 shows a comparison between the proposed algorithm and the Dynamic Bit Arbitration (DBA) [17]. Note that, our proposed method provide better performance, with both random and sequential ID's, even with large number of tags. Figure 5.6 shows the identification time of the proposed PBS algorithm with 40kb/s bit rate and ID length =16 bits. In this Figure, the identification time of the proposed algorithm is compared with the algorithms in [16]. Note that the BIBD technique [16] requires at least 450 ms for identification of 300 tags and 2550 ms is required by the traditional query tree algorithm to identify the same number of tags. The dynamic query tree algorithm identifies the same number of tags in 1500 ms. However, the proposed protocol can identify the same number of tags within 120 ms for random ID’s and 19 ms for sequential ID’s.

98

CHAPTER 5

PARALLEL BINARY SPLITTING PROTOCOLS

Fig.5.5 Total transferred bits vs. the number of the tags (32 bits long for all IDs)

Fig.5.6 Identification time of different quantity of tags under 40 kb/s bit rate (ID length =16 bits)

99

CHAPTER 5

PARALLEL BINARY SPLITTING PROTOCOLS

The proposed scheme can be summarized as follows: *The proposed protocol is based mainly on self assigning relative dynamic virtual orders, which regulate transmission order (tags responses) in a dynamically changed order queue. Tags are provided with the intelligence such that becoming able to adapt their relative order with respect to surrounding tags, that each tag continually changes its assumed relative position in virtual reply queue, by, first, each tag in the reply queue waits its reply order to successively send the next level bit ID, then receiving one bit feedback reader information informing tags with the result of each tag reply (collision or no collision). * Collision condition will be considered as splitting point that increases the overall number of appearing subgroups (branching check points). That point has two children in the binary tree to be checked in the next level, hence modulating its relative order of replying in the next level. * We can treat the tag ID as the tag address (the splitting rule) in the same time as the message to be transferred. The collision gives rise to increases the branches of the checked points, and remaps the pretended tree configuration. Hence, it does not need retransmission. By the end of splitting you get identification of all existing tags. It overcomes some methods that need retransmission on collision condition. *While “no collision” state has no change in relative tags order and overall regarded splitting subgroups. * Owing to order based splitting, Reader and tags (arranged in the virtual reply queue) agree on the current position of checked node in the binary tree successively without the need of transmitting the position of the currently transmitted bit. So, it performs better than the methods that required the restart node position.

100

CHAPTER 5

PARALLEL BINARY SPLITTING PROTOCOLS

* Minimizing transmission overhead of the needed data that reader must transmit to tags (minimized reader code), that will be confirmed in one information (one for collision, zero for no collision) to have faster identification. Each tag interprets this received information to modify its reply order, that its position in the reply queue will be changed accordingly in the next bit level transmission. *Simple tag reconstruction that needs two counters plus two registers operating under simple logic design. Tags have to be very simple in order to be as cheap as possible. *faster re-identification by the relative tags’ orders, that each tag saves its relative order , that reader can rapidly polling out tag existence ,because each knows its order to reply, without collision, so it improves reading rate in the subsequent Read Cycles. * The major advantages of the proposed scheme are low implementation complexity and lower number of bits transferred between the reader and the tags in the identification process. It consumes a number of bits equal to the twice of the number of binary tree node of existing tags. The performance analysis shows that the proposed technique outperforms most of the recent techniques in most cases.

101

CHAPTER 5

PARALLEL BINARY SPLITTING PROTOCOLS

5.3 Fast Parallel Binary Splitting Tree Anti-collision Technique (FPBS) The normal identification process needs successively exchanging messages between the reader and tags, where the next reply of one side depends on the other side. The tags will reply or stay silent according to the reader reply. Hence, each side waits the other side message before performing the next action. The overall identification time is the sum of the time consumed of the two sides. Although, the PBS path minimizes the dialog between the reader and tags to only one bit tag response followed by one bit reader reply, but the reader one bit response for each one bit of tags' reply is high overhead that must be overcome. The one bit reader response is used for reporting the collision state (1 for collision, 0 for no collision). This overhead information is clear from the result of the number of exchanged bits. The number of exchanged bits is equal to twice the number of the binary tree nodes of the existing tags except the leaves nodes. In this Section, we shall develop a fast tree-based protocol based on parallel binary splitting (PBS) technique to completely identify all the tags in the interrogation zone with minimum number of exchanged bits. The Fast PBS further minimizes the exchanged bits by confining the need of sending reader report to the collision condition only. The proposed algorithm aims at reducing the number of transferred bits between the reader and the tags to improve the characteristics of the reading rate and the identification speed. In PBS, the message length is one bit per message for both the reader and tags. It is mainly depending on exchanging one bit sequentially between the tags and the reader (one-to-one bit dialog).

102

CHAPTER 5

PARALLEL BINARY SPLITTING PROTOCOLS

Depending on the collision report received from the reader, each tag continually changes its relative position in the virtual replying queue. The tag used in PBS can communicate with half duplex ability. The complete description of PBS is described in the previous section.

5.3.1 Collision tracking assumption in FPBS (1) The reader can analyze the response of tags clearly and detect collision by using the Manchester code characteristics. (2) The reader checks whether a collision occurs or not in each bit on the received sequences. (3) The reader does not need to receive any data after receiving the first collided bit. (4) Reader can truncate unnecessary data bits to reduce the receiving time. (5) The reader transmits an ACK signal to stop tags transmission if there is a collision. It is a practical assumption and is defined in [23, [25], and [43]. These assumptions imply an increasing cost of full duplex communication tag ability.

5.3.2 FPBS OPERATION According to the PBS path, the “No Collision” state does not require any modifications in the assigned tags relative orders. It is only the collision state that requires changing in the relative orders and the control counters. Here, we shall assume that the “No Collision” is the default state. The tags will be allowed to send their marked bit in its previously assigned orders until receiving reader acknowledge notifying a collision state. It is not necessary for the reader to comment each transmitted bit from tags. If a data collision takes place, the reader sends acknowledge, else, no reader action is assumed. 103

CHAPTER 5

PARALLEL BINARY SPLITTING PROTOCOLS

By receiving the collision acknowledge from the reader (through phase or amplitude change in the reader continuous RF signal), all tags modify their next replying orders in the next bit level, without any change in the current replying orders in the current bit level. The new self assigned replying orders will be taken into account in the next bit level. As soon as receiving reader's collision ACK, all tags will stop transmission (any tag transmission could be discarded during the reader ACK) and change its control counters accordingly. Figure 5.7 shows the time diagrams of both PBS and FPBS. The collision bit does not need to be transmitted again. It leads to modifications in the contents of the internal counters of all tags involved in the identification session.

Fig.5.7 FPBS time diagram: (a) PBS, (b) FPBS

As shown in figure 5.7.b, the tags continue in transmitting in its predefined orders without reader interruption until the reader detects collided bit.

104

CHAPTER 5

PARALLEL BINARY SPLITTING PROTOCOLS

Hence the reader sends ACK to inform the tags that it is the time to change the orders (because of the insertion of a new subgroup). The system is assumed with full duplex ability. By that cost, it can reduce the identification time to the half of that in PBS without increasing the bit rate or the band width.

5.3.3 Performance analysis In this Section, the performance of FPBS algorithm will be discussed by two demonstration examples. Figure 5.8 shows the diagram of four tags in binary tree, as an example, {A, B, C, D} = {0000,0110,1110,1111} to be identified with each node contains its order of replying according the PBS path. Figure 5.9 shows the transmitted bit stream between reader and tags, for PBS and FPBS.

Fig.5.8 The diagram of four tags arranged in binary tree

Table 5.2 the anti-collision process of the proposed FPBS algorithm (--: silent)

105

CHAPTER 5

PARALLEL BINARY SPLITTING PROTOCOLS

Fig.5.9 The transmitted bit stream between reader and tags for parallel splitting algorithm by PBS & FPBS

The binary tree has 9 internal nodes (does not include the leaves). The reader reconstructs the binary tree by scanning the tree nodes as shown in Figure 5.4 and Table 5.2, and receiving tags response at each node. The table shows the reader action in the case of tags' collisions. In this example, the performance analysis is organized in a comparison with the recent algorithms as follows: • In the EAA algorithm [23], the total number of feedback bits and the reader response is 19 and 11; respectively. It uses 30 bits to identify the same number of tree nodes in our example. (19+11=30). •

In the proposed PBS algorithm, it consumes one bit for each node (tag response), and one bit for reader to report the state (collision or no collision).

The tree has 9 nodes (except leaves). The number of

106

CHAPTER 5

PARALLEL BINARY SPLITTING PROTOCOLS

transmitted bits by the tag is equal to the transmitted bits by the reader and is equal to 9 bit. Then, the overall transferred bits =18 bits. • In the FPBS algorithm, the number of transmitted bits by tag is 9 bits and the number of collision nodes is 3 nodes which equals the number of transmitted bits by the reader. Then, the identification time can be estimated as the time of transferring 9+3=12bits. It is clear that it has lower data overhead.

* In NEAA protocol [43], an example of identifying seven tags A, B, C, D, E, F, and G in the interrogating zone is proposed. Their tag IDs are “0000”, “0001”, “0010”, “0110”, “1001”, “1010”, and “1110”, respectively as shown in Figure 5.10.

Fig.5.10 the diagram of seven tags arranged in binary tree

• According to NEAA protocol, the total number of transmitted bits between tags and reader are 23+19=42 bits. The total number of feedback bits and the total number of response bits are 23 and 19, respectively. [43] •

However, in FPBS, it needs 13 bits as tags response, and 6 bits as reader collision ACK (note that: number of tags = number of collision nodes + 1). Hence, the total exchanged bits are equal to 13+6=19 bits.

107

CHAPTER 5

PARALLEL BINARY SPLITTING PROTOCOLS

5.3.3.1 The upper limit (bound) of the exchanged bits between the reader and tags *The number of transmitted bits from the tags equals the number of the nodes in the binary tree. *The number of transmitted bits from the reader equals the number of the collisions in the binary tree. The number of tags in the binary tree equals the number of collisions plus one. Hence, the number of reader transmitted bits (collision reports) is approximately equal to the number of identified tags. *The total number of exchanged bits = number of tree nodes (except leaves) + number of tags (tree leaves) = number of the binary tree nodes including leaves. The identification time of the proposed technique is the time required for sending number of bits equal to the number of the binary tree nodes of the existing tags. The upper limit (L) is considered the maximum number of exchanged bits between tags and its reader (interrogator). It must be less than the number of existing tags multiplied by the length of the Tag ID. For example: if you have at random 500 tag (with the tag ID length = 96 bit). Then, • The number of the exchanged bits must be less than 500*96 =48000 bits. However, under the bit rate of 80 kb/s, we need 0.6 second to complete the identification process in its worst case. • The number of the binary tree nodes is 43532 bits (without leaves). It has, as maximum, 500 collisions. Hence, the total number of exchanged bits is equal to 44032 bits (the total number of tree nodes including leaves). It can be identified in 0.55 second under the bit rate 80kb/s, which is smaller than the upper limit (0.6 sec.). 108

CHAPTER 5

PARALLEL BINARY SPLITTING PROTOCOLS

5.3.4 Simulation results

Fig.5.11. Total transferred bits vs. the number of tags, for random IDs.

Figure 5.11 and Table 5.3 show comparison among the proposed (FPBS) algorithm, the parallel binary splitting (PBS) and the Dynamic Bit Arbitration (DBA) [17]. The number of tags is increased from 50 to 500 tags. Here, the simulation results of these three algorithms is done to study the total transferred bits in identification of different number of tags, for random IDs (32-bit long for all IDs). From this comparison results, it is clear that the proposed FPBS protocol provides better performance than the other two approaches. For example, to identify 250 tags using DBA method, 14060 bits are required in the identification process. While, for the same number of tags, PBS method transfers 12014 bits. However, to identify the same number of tags using the proposed FPBS algorithm, only 6007 bits is required, which represents the half of the required bits for PBS method.

109

CHAPTER 5

PARALLEL BINARY SPLITTING PROTOCOLS

Table.5.3 Estimated number of exchanged bits (for DBA, PBS, and FPBS)

no. of tags 50 100 150 200 250 300 350 400 450 500

DBA 3024 5894 8638 11392 14060 16726 19332 21952 24478 27186

Number of exchanged bits PBS FPBS 2650 2700 5098 5198 7458 7608 9664 9864 12014 12264 14328 14628 16572 16922 18800 19200 20982 21432 23138 23638

Figure 5.12 shows the average number of identified tags per second under the following simulation conditions: In the field of the reader, the number of tags is increased from 2 to 512 and the length of the tag IDs is 96 bits. Both tag-to-reader data-rate and readerto-tag data-rate are chosen to be 80 kbps. By generating 900 tags (with ID long =96 bit) randomly, the binary tree of the FPBS protocol will have 77598 internal nodes which equals the number of tags transmitted bits. The reader will send 900 bit as collision reports. The total exchanged bits between the reader and tags are 78498 bits. It consumes 0.9812 second under bit rate of 80kb/s. As shown in this Figure, the FPBS provides the best performance. Figure 5.13 shows the average required overhead-bit for one tag identification (as the cost due to prefix and iteration overhead) under the same simulation conditions. On the average, the reader overhead is equal to one bit per tag. The proposed FPBS consumes the least reader overhead per one tag identification. This Figure confirms one of the advantages in the proposed FPBS algorithm.

110

CHAPTER 5

PARALLEL BINARY SPLITTING PROTOCOLS

Fig. 5.12 The average number of identified tags per second

Fig.5.13 Search cost (Bits): average required bits for one-tag identification

111

CHAPTER 5

PARALLEL BINARY SPLITTING PROTOCOLS

Fig.5.14. Comparing average identifying time of the proposed against recent algorithms

Figure 5.14 shows the results of identifying 256 tags with variable ID length from 8-bit to 64-bit. By assuming that the time of transmitting one bit is 5 µs, then, if the proposed FPBS protocol is used to identify the 100% tag density of 8 bit ID length (i.e.2^8=256 tags are existing in the interrogator’s operating range), the reconstructed binary tree has 255 internal nodes (without including the tree leaves). It needs 1 bit tags' reply at each node plus 255 collision reader report. There is a collision in each node. The equivalent number of exchanged bits is 510 bit.

The total

identification time =510* 5 µs= 2550 µs. It provides 10 µs per one-tag identification according the proposed FPBS.

However, the NEAA

algorithm in [43] consumes 16.8 µs to identify one tag.

112

CHAPTER 5

PARALLEL BINARY SPLITTING PROTOCOLS

5.3.5 FPBS Performance in successive sessions In this section, the performance of the anti-collision algorithm will be explored in the case of some tags are arriving or leaving in successive reading cycles. The proposed FPBS can provide faster performance in the first and the successive reading rounds. Figure 5.15 shows the example of the binary tree of the four tags which was tested in the performance analysis.

Fig.5.15 the identified tree in the last reading session

Each tag saves its relative order in the current order register (COR). The total number of the identified tags can be found in the current path register (CPR). For example, tag “C” has the COR=3. All tags have CPR=4 which is the total count of identified tags at the session end. • To take the benefits of the order based FPBS protocol, we can define two types of reader sessions: 1- Checking session: the reader checks the existence of the recognized tags in the last session. The reader starting command contains the total count of the identified tags “CPR”. All tags that have the same “CPR” will respond with one bit in its relative order (allocated time slot) to inform the reader with its existence. The reader and tags are operating in simple one-to-one dialog in a frame of

113

CHAPTER 5

PARALLEL BINARY SPLITTING PROTOCOLS

CPR time slots. For example, if the reader starts checking session for the tree shown in Figure 5.15, the “CPR” represents a frame with four time slots. Each tag should respond by one bit in its time slot. If the tag responds to the reader, the reader replies by one bit ACK = “0” to say “no change in the relative orders”. But if tag “C” for example went out the reader range, its allocated time slot will be idle, and the reader will send one bit ACK = “1” to say “there is a leaving tag in this time slot, and the remaining tags must update its relative orders accordingly”. Tag “D” will modify it’s COR to be 3 instead of 4. All staying tags (A, B, D) will modify its CPR register to be 3. In general, it reduces the need of transmitting the 96 bit of every tag ID, into one bit response representing its existence in its allocated time slot.

2- PBS reading session: the reader command starts the parallel splitting in the binary tree for the newly arriving tags. * If the command contains “CPR”=1, then all tags will reset its counter and registers, and operate in the normal parallel splitting (PBS). *If the command contains “CPR” > 1, then: - The tags with the same “CPR” will remain silent, because they were recognized in the previous session. Hence, we can prevent staying tags from colliding with arriving tags. Staying tags will listen to the reader responses in the splitting process of the arriving tags, and update their “CPR”, such that updating the total number of tags in the reader range. The “CPR” will be incremented at every reader collision report. The new “CPR”= the old “CPR”+ number of reader collision reports+1. - The tags that have different “CPR” will reset their counter and registers, and operate in the normal parallel splitting. But, at the end of the session, it will add the received “CPR” from the reader command to both “CPR” and 114

CHAPTER 5

PARALLEL BINARY SPLITTING PROTOCOLS

“COR” registers. This action will combine the two subgroups of the staying tags and the arriving tags into one recognized group that can be checked in successive reading sessions.

Fig.5.16. the new tree in the next reading session

Suppose that there are two arriving tags (F=0,1,0,0 and G=0,1,0,1) for example as shown in Figure 5.16. The reader sends a command with the previous “CPR”=4, then the four Tags (A,B,C,D) will remain silent and listen to the number of reader collision reports. Tags (F and G) have one collision node; hence all tags will update its CPR to be 6. The relative orders (COR) of the two tags F and G will be 5 and 6.

5.3.6 Concluding remarks of FPBS protocol FPBS protocol presents an enhanced version of the PBS approach for tag anti-collision. The proposed algorithm overcomes the prefix and iterations overhead of the previously proposed protocols in literatures. The increased cost is the assumption of using full duplex tag which is practical assumption. The tag achieves the self transmission control by aid of two counters and registers. The information needed for tag operation is the collision notification which can be provided by reader. The tag will modify its replying order in the next bit level according to the collision condition. 115

CHAPTER 5

PARALLEL BINARY SPLITTING PROTOCOLS

The major advantages of the proposed scheme are the low implementation complexity and the minimum number of transferred bits between the reader and the tags in the identification process without increasing the band width or the bit rate. It consumes one bit per tag as a reader overhead. So, the total number of exchanged bits is equal to the number of the binary tree nodes (including leaves) of the existing tags. It also provides faster performance in the successive reading cycles. It can check the existence of the previously recognized (staying) tags very fast. It also prevents the staying tags from collision with the arriving tags with minimum overhead. The performance analysis shows that the proposed technique outperforms most of the recent techniques in most cases. • In the next section, another modification on the PBS protocol will be suggested to develop an integrated anti-collision protocol for multitag multi-reader environment.

5.4 Integrated Reader and Tag Anti-collision Protocol based on Similar Topology Trees (STT) in the one collision domain Although both collisions modes (tags and readers collisions) increase the identification delay in Radio Frequency Identification (RFID) system, all previous work discussed the tag and the reader collision problem separately. In this section, a new protocol is introduced for both tags and readers anti-collision in multi-tag multi-reader environment. The proposed anti-collision protocol is based on constructing similar topology binary trees for all readers in the collision domain. The new Parallel Binary Splitting (PBS) identification path will be used in this process. All readers are synchronizing to reply at the same time with the similar information to build similar binary trees topology. The identification time is divided

116

CHAPTER 5

PARALLEL BINARY SPLITTING PROTOCOLS

among the tags and the readers. One bit reply will be sent sequentially (one-to-one bit dialog) in the identification process. All tags lie in the overlapped reader-region will be identified one time only. The integrated treatment of the collision problem will provide minimum exchanged bits with minimum overhead, higher throughput, and simple logic operation. The main idea of the proposed algorithm is based on applying the PBS searching path on multi-reader environment with minor rectifications in reader side, and without any tag modification. The proposed method provides an integrated and fast solution for solving both readers and tags collision simultaneously in multi-reader environment. The reader provides the tags one bit comment about the collision state of tags (collision or no collision).

5.4.1 Similar Topology Trees (STT) Even though many anti-collision protocols for tags collision and reader’s collision has been proposed, we feel that the integration protocol for both collision types have not been exploited in reducing the identification time while maintaining a reasonable transmitting bits. The STT protocol does not require smart tags to detect collision. It applies the PBS technique to build similar binary trees according to the new identification path of the PBS, in solving both collision problems. In STT, virtual (silent) paths are added to preserve the similar tree configuration for all readers. The proposed similar topology tree (STT) anti-collision protocol was built to solve the data collision problem even with the following collision sources: (1)

simultaneous

tags

transmissions,

(2)

simultaneous

reader’s

transmissions, (3) simultaneous tags and readers’ transmissions. Any anticollision protocol must resolve these data collision sources in the identification process. The proposed anti-collision protocol is described next with some inherent assumptions. 117

CHAPTER 5

PARALLEL BINARY SPLITTING PROTOCOLS

5.4.2 Practical assumptions of STT protocol In the proposed anti-collision algorithm, the assumptions are considered as follows: (1) The position of the readers and the tags are considered in one collision domain, (2) Tags in active state send their marked bit simultaneously. The role of the reader operation is to detect the collision and to provide tags with collision report to dynamically update their replying orders (in the next bit level), (3) According to the PBS technique, the time is divided between tags and readers; each side will send one bit reply sequentially (one-to-one bit dialog). There is no chance for simultaneous tags and readers’ transmissions. This solves the third problem of simultaneous tags and reader’s transmissions, (4) To solve the second collision problem (collision due to simultaneous reader’s transmissions), all readers must be synchronized (forced) via the control channel between the readers to send a similar reply simultaneously, (5) Similar binary trees topology for all readers are constructed, in one collision domain, with one collision bit step. This means that all readers will send the same reply (1 or 0) to represent the collision state of last transmitting bit from the tags. Forcing all readers to send the same reply will provide some silent paths in some reader’s tree to keep similar configuration of all readers, (6) The reader communicates with each other to report a collision state if any reader detects a collision. Readers share its one bit collision state (collision or no collision) through a separate control channel, (7)

The tags can

automatically change the contents of their internal counters and registers to represent the new relative replying order in the next bit level.

118

CHAPTER 5

PARALLEL BINARY SPLITTING PROTOCOLS

5.4.3 Reader and tag operation based on Similar Topology Tree (STT) Our protocol is based on remapping the discovered binary tree configuration after each tag and reader one bit dialog. * For the tags, the remapping process will be done relative to the contents of two counters and two registers. By knowing the state of the last tags reply, from reader report, each tag continually changes its relative position in the virtual replying queue, exactly as in PBS without any modifications. * However for the readers, virtual PBS path will be generated when the reader receives the collision report from another reader outside its detecting range due to sharing collision state of each tag's one bit transmission through the control channel to build similar binary trees for all interrogating readers. Figure 5.17 shows the STT Protocol flowchart. * The Reader and tag operation can be summarized as follows: 1- All readers will be synchronized to start interrogation session. 2- Initially, each tag starts by thinking that it is the only tag in the reader range and resets the two counters and the two registers. Then, it will send the first marked bit. 3- All readers are listening to tags bits. Then, they are continually updating the pretended binary tree of existing tags, in their reading range, according to tags reply. 4- All readers share the received bit from each reader in its range through the control channel to send the same reply as follows: a-

If the reader detects a tag collision i.e. receiving one and zero at the

same time, then informing all readers to send one bit collision report to build similar topology tree (PBS tree). In this case, the inserted path will be

119

CHAPTER 5

PARALLEL BINARY SPLITTING PROTOCOLS

considered as: true path for the readers that detect collision in their ranges, and virtual path for the readers that detect no collisions in their ranges. b-

If the reader doesn’t detect tag collision in its interrogation range, it

listens to other readers. •

If any reader detects collision, all readers will follow it by sending

collision report in its reading range. In this case, the inserted path will be virtual path (silent path) to keep similar tree topology for all readers. Note that, the inserted silent paths is coming from tags lying outside its range. •

Else, all readers send no collision report.

5-

Tags update their replying order of the next marked bit according to

the received readers’ common report. 6-

The readers will listen to the next marked bits.

7-

If it is not the last bit, then go to step 4-a, Else, End.

Fig.5.17. STT Protocol flowchart

120

CHAPTER 5

PARALLEL BINARY SPLITTING PROTOCOLS

5.4.4 Performance Analysis In this section, the performance of STT protocol will be discussed. The idea of similar topology trees will be clarified through the next example. Figure 5.18 shows a distribution of six tags and three overlapped readers, as an example, {T1, T2, T3, T4, T5, T6} = {010010, 010101, 010110, 011010, 011000, 011011} to be identified.

Fig.5.18 Distribution of three intersected readers with six tags.

We have three overlapped reader’s region (R1, R2, R3). Each reader builds its binary tree under the proposed protocol. Figure 5.19 shows the complete identified binary tree of the six tags relative to one reader covering the whole tags. However, Figure 5.20 shows the similar topology trees of the readers and the existing six tags. All readers will share its collision state through a control channel.

121

CHAPTER 5

PARALLEL BINARY SPLITTING PROTOCOLS

Fig.5.19 The complete binary tree of the existing six tags

Fig.5.20 Three readers are building similar trees

122

CHAPTER 5

PARALLEL BINARY SPLITTING PROTOCOLS

When the six tags send their first ID bit (“0”) the readers will report no collision state to all tags and the identification paths will not increase. This case will be repeated with the second ID bits of the tags. There is a collision state when the tags send its third and fourth bits in all readers’ ranges. The splitting will be identical for the three readers. However, there are collisions in some readers’ ranges at the fifth bit, in this case, a collision report will be sent to all tags and new virtual (silent) path will be generated for some readers that detecting no collision. Virtual nodes (paths) in the reader’s binary tree are corresponding to tags outside the reader range. It is important to note that, Virtual (silent) paths are added to preserve the similar tree configuration for all readers as shown in Figure 5.19. For example, T6 lies in reading range of R3. Then, there is a true path with respect to reader R3, while there is a virtual path with respect to R1 and R2. Moreover, as shown in Figure 5.19, Reader R1 has two silent paths due to the transmitting bits of tags T5 and T6. The worst case of the identification time can be estimated as the time consumed by the PBS protocol for identifying all tags with one reader. For the proposed protocol, the identification time is independent of the number of existing readers. Tags in overlapped readers’ ranges will be identified one time with all readers. For example, although T1 lies in the reading ranges of the three readers, it will read by all readers once at the same time because all readers work on parallel. The upper limit (worst case) of the exchanged bits for PBS path • Worst case (W): is the maximum number of exchanged bits among tags and their interrogator. W < 2 * number of existing tags * number of bits of Tag ID

123

CHAPTER 5

PARALLEL BINARY SPLITTING PROTOCOLS

• For example if you have at random 500 tag (with the tag ID length = 96 bit) then, Number of exchanged bits must be less than 2*500*96 =96000 bit. • Under bit rate of 100 kb/s we need, as maximum, 1 second to complete identification.

5.4.5 Simulation Results Simulation is carried out to study the performance of the proposed protocol in a comparison form with some of recent protocols. The identification time and the system throughput of the proposed protocol over conventional reader anti-collision algorithms are studied in this Section. A case of identifying 500 tags with different number of readers (from 1 to 12) is tested. • Similar topology tree (STT) protocol is estimated under its worst case, i.e. with the maximum number of exchanged bits. The maximum number of exchanged bits is considered as twice the number of existing tags multiplied by the number of ID bits per tag. • Max. Exchanged bits < 2*500*96 bit = 96000 bit = 12000 byte. • With bit rate of 100Kb/s: The 96000 bits are transmitted within one second. This provides a network throughput of nearly 12000 bytes/second. • With bit rate of 50Kb/s: The 96000 bits are transmitted within two seconds. This provides a network throughput of 6000 bytes/second. Figure 5.21 shows the results of the network throughput when the proposed STT algorithm and five of recent algorithms are used to identify the current 500 tags. This Figure confirms one of the advantages of the proposed protocol.

124

CHAPTER 5

PARALLEL BINARY SPLITTING PROTOCOLS

Fig.5.21 Comparison of the network throughput at 50 kb/s bit rate

For the same number of tags, the proposed protocol results in constant identification time independent of the number of readers as shown in Figure 5.22.

Fig.5.22 Comparison of the identification time

This performance is achieved due to the synchronization of all readers to build the same complete binary tree. Although the proposed protocol work under one collision domain, the work with one collision bit is resolving. It takes benefits of the simple one-to-one bit dialog of the parallel binary 125

CHAPTER 5

PARALLEL BINARY SPLITTING PROTOCOLS

splitting identification path, without retransmitting any information in the collision state. As mentioned, most of the previous protocols utilize the TDMA scheduling. So, the main drawback of these protocols is the long waiting time and the large overhead information. Moreover, the tags in the overlapping reading ranges will be identified several times. However, in the proposed STT protocol, the tags lying in the reading ranges of several readers will be identified one time only.

5.4.6 Concluding remarks of STT anti-collision protocol This chapter presents a new integrated tag and reader approach for anticollision in RFID systems. The proposed protocol is based on constructing similar topology binary trees for all readers in one collision domain. It is based on a parallel tree-scan (PBS) with the self modified relative order of tags. All readers are synchronizing to send (at the same time) the same reply and to build similar topology trees (STT). The proposed protocol provides constant identification time independent of the number of existing readers. The major advantages of the proposed scheme are low implementation complexity, the integrated solution of collision problem, and lower number of transferred bits between the readers and the tags in the identification process. Moreover, due to the simple dialog between readers and tags, the proposed protocol provides the best system throughput relative to the recent anti-collision protocols. The main advantage of the proposed protocol arises from the integrated treatment of the two collision problem of the readers and the tags. The integration of tags and readers anti-collision mechanism provides minimum exchanged bits with minimum overhead and simple logic operation.

126

CHAPTER 6 ANTI-COLLISION PROTOCOL BASED ON UP-DOWN COUNTER AND ONE BIT READER RESPONSE Abstract This chapter describes a simple and fast counter based anticollision protocol. The advantages (simple one depth counter in tag side & simple dialog & one bit reader response at collision or identification & faster than the traditional counter based anti collision protocols) of the protocol will be clarified. Then the solution for both stationary and moving scenarios is suggested with minor modification in tag side by adding additional counter.

CHAPTER 6

UP-DOWN COUNTER AND ONE BIT READER RESPONSE

CHAPTER 6 ANTI-COLLISION PROTOCOL BASED ON UP-DOWN COUNTER AND ONE BIT READER RESPONSE 6.1 Introduction This thesis addressed the problem of data collision in multi-tag RFID system. The main criteria used in measuring our contribution is to develop fast protocol which consumes minimum number of exchanged bits between the reader and tags, with the simple logic needed in tag side. This chapter describes a simple and fast counter based protocol. The tag processing is depending on one depth counter and conflict pointer to determine its position from the identification state. The main advantage of the proposed protocol is the simple one bit reader response in the identification process. The one bit reader response will provide the tags all information about the new state of the current depth position of replying. The tags will transmit their replying order without reader interruption until the reader detects the collision state or the identification state. Based on this idea, the identification process will be faster than the traditional counter based anti collision protocols and the overhead information will be reduced.

The main drawbacks that we will try to overcome in that protocol are: 1- Long reader command. 2- Complex logic required on tag side. 3- The overhead information required to inform the tag about the collided bit position in the collision code. It is required to build simple and fast anti-collision protocol that avoids the drawbacks of the previous counter based protocols.

127

CHAPTER 6

UP-DOWN COUNTER AND ONE BIT READER RESPONSE

6.2 Protocol Operation Assuming that inside each tag there is a depth counter and collision pointer. The depth counter (D) will measure how far the tag position from the current replying tags. The collision pointer (P) will keep tracking the marked bit that will be transmitted when being in the active replying state. The tags will transmit the ID bits without reader interruption until the reader detects collision or identification.

6.2.1 Reader Operation The reader starts the communication session by transmitting the continuous RF signal to power the passive tags. Then, all tags will be in the active state and the ID, bit by bit, will be transmitted. The reader then receives the tags responses until the collision state (at collision node) or finishing the identification path (at leaf node), i.e. the tags will transmit their IDs without reader interruption until the reader detects collision or identification. The reader one bit response (1 for collision and 0 for identification) will be sent to inform the tags how to change their current replying depth to change their state. If the reader detects a collision state at the end of any path in the binary tree, it will be treated as two tag identification. Hence, it sends the identification response. Then, the reader will return to the nearest waiting tags whose depth counter becoming zero.

6.2.2 Tag Operation All tags will start in the active state as response to the reader command. The current active tags will continue replying their ID bit by bit until completing identification or receiving a collision notification from the reader.

128

CHAPTER 6

UP-DOWN COUNTER AND ONE BIT READER RESPONSE

Fig.6.1 State diagram of the proposed up-down counter protocol

In the case of receiving the collision notification, there are two possibilities: 1) the active tags replying 0 will continue in the active state without affecting the depth counter, 2) the active tags replying 1 will stop transmission and increment their depth counter. Then the tags will be in the waiting state. It stays silent and the reader will continue in the path of the active tags. It is important to note that, the depth counter (D) for the tags in the waiting state will be incremented with the collision acknowledgment, and will be decremented with the identification acknowledgment. Also, the depth counter for the waiting tags will be decremented by the reader identification response until zero count (D=0), then the state will be changed to the active state. It is not necessary to tell the tag any information about the position of the starting node. The information about the last bit shifted out is determined by the collision pointer. The tag will be in the identification state when it has sent all its ID bits. Figure 6.1 shows the state diagram of the proposed protocol.

129

CHAPTER 6

UP-DOWN COUNTER AND ONE BIT READER RESPONSE

6.3 Performance Analysis 6.3.1 Demonstration example

Fig.6.2 The diagram of four tags binary tree to be identified

In this section, the performance of the proposed algorithm will be discussed. Figure 6.2 shows the diagram of four tags in binary tree, as an example, {A, B, C, D} = {0000, 0110, 1110, 1111} to be identified with each node contains its replying order according to the Depth First Search (DFS) path. It is clear in this figure that the reader needs to hear one bit reply of the interrogated tags at each node. Tags and the reader agree to follow the interrogation order of these nodes according to the proposed protocol. Reader reconstructs the binary tree by scanning the tree nodes in the shown order in Figure 6.3. This Figure describes in details the process of node exploration under the control of its depth counter. The contents of depth counter are shown also in this Figure after each step. Figure 6.4 shows the bit-exchanged stream between the reader and tags during the identification session. As mentioned, the reader responds only to inform the tags with the branching moment in the collision state or the return action in the identification state.

130

CHAPTER 6

UP-DOWN COUNTER AND ONE BIT READER RESPONSE

Fig.6.3 The order of node exploration

131

UP-DOWN COUNTER AND ONE BIT READER RESPONSE

CHAPTER 6

Fig.6.4 Transmitted bit stream between the reader and tags

In this example, the performance analysis of the proposed algorithm is organized in a comparison form with the recent algorithms as follows: For the number of tree nodes (except the leaves node) = 9 nodes. • In the EAA algorithm [23], the total number of feedback bits and the reader response is 19 and 11; respectively. It uses 30 bits to identify the tree nodes in our example. (19+11=30 bit). • In the PBS algorithm , it consumes one bit for each node (tag response), and one bit for reader to report the state (collision or no collision). The number of transmitted bits by the tag = the number of transmitted bits by the reader = 9 bit. Then, the overall transferred bit =18 bits. • In the proposed algorithm, the binary tree has 9 internal nodes (the tree nodes except the leaves), hence, it consumes 9 bit as tags response. It consumes 5 bit as reader response (2 collision notification + 3 identification notification). As mentioned, the collision at leaf node will be replied as identification response. The overall transferred bit between the reader and tags = 14 bits. This improvement will be more sensed by increasing the number of identified tags and the number of bits per tag. 132

UP-DOWN COUNTER AND ONE BIT READER RESPONSE

CHAPTER 6

6.3.2 Another example In [43], an example of identifying seven tags A, B, C, D, E, F, and G in the interrogating zone is proposed. Their tag IDs are “0000”, “0001”, “0010”, “0110”, “1001”, “1010”, and “1110”, respectively. The total number of feedback bits and the total number of response bits are 23 and 19, respectively. The total number of transmitted bits between tags and reader are 23+19=42 bits [43]. Collision at their last node means: two tag

Fig.6.5 example of seven tag binary tree identification.

However by the proposed protocol, only 13+11=24 bits are used to identify the same number of tags. As shown in Figure 6.5, the constructed binary tree of the seven tags has 13 internal nodes. So, the tags will reply by 13 bits and the reader will send 5 bits for collision notification and 6 bits for identification notification. Figure 6.6 and table 6.1 compares the proposed protocols in the seven tag identification example. Table 6.1 Exchanged bits compare

Fig.6.6 protocol compare in seven tag identification example

133

Total exchanged bits

Protocol

transmitted bits by Tags

transmitted bits by Reader

NEAA [2010]

23

19

42

PBS

13

13

26

Up-Down Depth Counter

13

11

24

FPBS

13

6

19

CHAPTER 6

UP-DOWN COUNTER AND ONE BIT READER RESPONSE

6.3.3 Data overhead per one tag identification If we consider the reader responses as the overhead information in the process of tag identification, the overhead in the proposed protocol can be considered two bits per one tag as maximum. Hence, in the proposed protocol: * The Max. Exchanged bits = 2 * number of existing tags + Number of the binary tree nodes except leaves. In general, if the tag has n bit length, then the overall time needed for one tag identification must be less than the time needed for the transmission of (n+2) bits, which is the upper bound of the transmitted bits per tag. However, the upper bound of the algorithm in [43] is computed relative to the number of time slots. It has four types of time slots with three bits instruction code, besides sending the position of the collided bit along with the collision code.

6.4 simulation results Figure 6.7 and Table 6.2 provide a comparison among Dynamic Bit Arbitration (DBA) in [17], Parallel Binary splitting (PBS), and the proposed algorithm by measuring the total transferred bits between the reader and tags (with 32 bit ID long). In the simulation results, to identify 250 tags, DBA uses 14060 bit while PBS uses 12014 bit. However, to identify the same number of tags using the proposed algorithm, only 6507 bit are required as the total exchanged bits, including 500 bits are the maximum reader overhead.

134

UP-DOWN COUNTER AND ONE BIT READER RESPONSE

CHAPTER 6

Fig. 6.7 Total transferred bits vs. the number of tags, for 32 bit ID long. Table 6.2 Estimated number of exchanged bits.

Random IDS (32 bit tag ID long) no. of tags

DBA [17]

PBS

Up-Down counter

50

3024

2650

1425

100

5894

5098

2749

150

8638

7458

4029

200

11392

9664

5232

250

14060

12014

6507

300

16726

14328

7764

350

19332

16572

8986

400

21952

18800

10200

450

24478

20982

11391

500

27186

23138

12569

Figure 6.8 shows the results of the identification time with respect to the number of tags when the proposed algorithm and four of recent algorithms are used with 40kb/s bit-rate and the ID length is 16 bits. In this Figure, for the BIBD technique [16], at least 450 ms is required to identify 300 tags

135

CHAPTER 6

UP-DOWN COUNTER AND ONE BIT READER RESPONSE

and 2550 ms is required to identify the same number of tags by the traditional query tree algorithm. The dynamic query tree algorithm identifies the same number of tags in 1500 ms. However; the proposed protocol can identify the same number of tags within 60 ms.

Fig.6.8 Identification Time of different number of tags with 40kb/s bit rate (ID length =16 bits)

Fig. 6.9 Average identifying time of each algorithm.

136

CHAPTER 6

UP-DOWN COUNTER AND ONE BIT READER RESPONSE

Figure 6.9 shows the results of identifying 256 tags with variable ID length from 8-bit to 64-bit. Assume that the time of transmitting one bit is 5 µs under bit rate of 200kb/s. Then, if the proposed protocol is used to identify 256 tags of 8 bit ID length (the 100% tag density of 8 bit ID length i.e.2^8=256 tags are exist in the interrogator’s operating range), the reconstructed binary tree has 255 internal nodes (without including the tree leaves). It needs 1 bit tags reply at each node. There is a collision in each node provides 1 bit reader collision response and 128 bit reader identification. The total number of transferred bits is 638 bit. The total identification time =638* 5 µs= 3190 µs. It provides 0.0124 ms per one tag. However, [43] consumes 0.0168 ms to identify one tag.

To demonstrate the performance of the proposed algorithm the following example is also tested.

Fig. 6.9 Average identifying time with various numbers of tags (Tag ID length=12bit)

137

CHAPTER 6

UP-DOWN COUNTER AND ONE BIT READER RESPONSE

In this example 4096 tags with 12-bit ID length are tested. It has 12- level full binary tree with 8183 nodes (including leaves). After applying the proposed Up-Down Depth Counter protocol, the following results are achieved: * The number of transmitted tags-bit is 4087 bits (internal tree nodes). * The number of transmitted reader-bits is 4087 bits, which are transmitted as collision notification, PLUS 2048 bits as identification notification. Hence, the reader transmits totally 6135 bits. *The total exchanged bits are equal to 10222 bits and the average exchanged bit per tag is less than 3 bit. It means that the average identification time is less than 0.015 ms per tag. Figure 6.9 shows the average identifying time relative to the number of existing tags starting from two tags to the full tree tags. It is clear from this figure that the proposed protocol identified the tags faster than the seven protocols used in this example.

6.5 Performance under moving scenario The relative movement between the reader and the tags leads to some tags are arriving or leaving the reader range. The current questions: • Can the protocol continue in the identification session in the case of moving tags or reader without disturbing the identification process? • Can the protocol exploit the obtained information from the last frame to prevent collisions between staying tags, and also avoid collision between arriving tags and staying tags for providing faster identification in successive interrogation sessions (frames)?

138

UP-DOWN COUNTER AND ONE BIT READER RESPONSE

CHAPTER 6

In Blocking ABS protocol (BA) [44], there are some answers to these questions. BA can be considered as a modified version of the Adaptive Binary Splitting (ABS) [22]. ABS and BA was studied in chapter 3 in Section 3.4.3.

6.5.1 Leaving tags The leaving (disappearing) tag go away from the reader range. It stops replying the reader. When the reader loses the communication received from the replying tag, it can continue to communicate with other remaining tags by sending (0) which stands for the identification message. That reply is used to inform other waiting tags that the reader finishes the advance in the followed identifying path by decrementing its depth counter. The leaving tag can be cleared from the binary tree by the same reader feedback used for the tag identification case. For example, in the Figure 6.11.a, if the tag (B) leaves the reader range, the reader detects that the current replying path is lost; hence the reader sends (0), as end the identification of that path. Then the reader and tags transfer to the nearest waiting tags (Path of C and D) as in Figure 6.11.b. Hence, leaving tags can be cleared from the reply queue by the reader identification notification.

Fig.6.11 The binary tree with the leaving tag (B)

139

CHAPTER 6

UP-DOWN COUNTER AND ONE BIT READER RESPONSE

6.5.2 Arriving Tags The arriving tag which enters the reader range will be powered and enter to the waiting state as shown in the state diagram in the Figure 6.1. The proposed protocol doesn’t allow the arriving tags to participate in the communication until starting the next interrogation session by receiving the start command to change its state to the active state. Then, all tags (staying and arriving) start the session equally. • The proposed protocol can work in the identification session with moving tags or reader without disturbing the identification process. The proposed protocol (in its simple one counter form) does not exploit the obtained information from the last frame. • Although BA [44] protocol can exploit information obtained from the last frame to achieve faster identification in successive interrogation sessions (frames), but the proposed Up-Down Counter protocol has the superior overall performance. The proposed protocol has the lowest reader overhead per tag (less than 2 bit per tag), but BA has the same drawbacks of [22], longer reader response, the random splitting rule and the more complex tag operation, hence larger overhead per tag. •

It is clear that, the proposed Up-Down Counter protocol has the superior overall performance due to the simple dialog based on one bit reader response and the fact that the collisions required, only, one bit reader collision ACK.



It is very important to note that, the proposed protocol can be modified to exploit the obtained information from the last frame to achieve faster identification in successive interrogation sessions. This can be done by adding another counter (Order Counter) to the

140

CHAPTER 6

UP-DOWN COUNTER AND ONE BIT READER RESPONSE

tag. The Order Counter (OC) tracks the relative order of the recognized tag. It will be incremented when the tag detects a collision ACK in its waiting state. Hence, each tag can remember its replying order (recognized depth) in the next rounds to avoid collisions among arriving and staying tags.

6.6 Conclusions This protocol presents simple and fast anti-collision algorithm. It overcomes the main drawbacks of counter based anti-collision algorithm. It depends on only one up-down counter controlled by simple logic of reader one bit response. It achieves great save in the reader overhead by using simple one bit response (1 for collision, 0 for identification). The tags will transmit its ID bit by bit without reader interruption until the reader detects collision or identification. The tag state will be changed from the waiting state to the active state without transmitting any information about the position of the restarting node. The simplicity of the proposed algorithm is due to the low cost of integration in the passive tag. Moreover, the proposed algorithm utilizes the identical prefixes IDs.

It works with

minimum overhead. Moreover, it can clear the leaving tags from the replying queue during the interrogation session without disturbing the advance in the identification process. The simulation results have shown that the collision recovery scheme is very fast and simple relative to the recent anti collision protocols in literatures. It works with minimum overhead in both stationary and moving scenario. It can exploit the information from the last frame to achieve faster identification in successive interrogation sessions using additional counter in tag side.

141

CHAPTER 7 CONCLUSIONS AND FUTURE WORK Abstract In this chapter, we conclude the proposed work, give a comparison between the proposed anti-collision protocols in this thesis, and suggest certain points of research such as hardware implementation, security issues and the problems of RFID tag insertion in banknotes.

CONCLUSIONS AND FUTURE WORK

CHAPTER 7

CHAPTER 7 CONCLUSIONS AND FUTURE WORK 7.1 Conclusions This thesis addressed the problem of data collision in RFID systems. In this work, we have presented anti-collision techniques that are used for resolving the collision problems with minimum exchanged bits between the reader and tags. The key idea was in building simple and fast dialog between the reader and tags to override the data overhead in the previous anti-collision protocols. The proposed protocols achieve improvements in the speed of the identification with the minimum required hardware in the passive tag. Our work provides a new identification path of parallel binary splitting (PBS) in the binary tree. It leads to one-to-one bit dialog between the reader and tags. Then, we provide fast enhanced PBS to reduce the exchanged bits, through the usage of full duplex tag, (FPBS) protocol. PBS protocol emphases the ability of fast identification in successive reading sessions. It also has the blocking effect which blocks collisions between arriving tags and staying tags in the successive sessions. Applying the PBS to the multi-reader environment produces an efficient and integrated solution to both collision problems through the proposed Similar Topology Tree (STT) with minor modifications in reader side. Also, we introduced an enhanced counter based anti-collision protocol with more simple operation than other counter based protocols of the first depth search path (FDS). It can work with the moving scenario without disturbance in the identification process. Generally, we believe that our frameworks are good candidates for fast anti-collision identification protocol for using in dense tag environment. Table 7.1 presents a comparison of the proposed anticollision protocols. 142

CONCLUSIONS AND FUTURE WORK

CHAPTER 7

Table 7.1 Comparison of the proposed anti-collision protocols Protocol

Identification speed

PBS

Medium

FPBS

The fastest

Full duplex tag with two counters and two registers

fast w.r.t all reader anticollision protocols

*Simple by using one collision domain & similar trees& *Half Duplex tag with two counters and two registers

STT

Up-Down Depth Counter Enhanced Up-Down Depth Counter

Very fast

Very fast

Tag Complexity

Designed for

Half Duplex tag with two counters and two registers

Tag anti-collision *Tag anti-collision *it can achieve faster identification in successive reading rounds

Both tags and readers anticollision

Very simple (one depth counter &Full duplex tag )

Tag anti-collision

simple (one depth counter & order counter)

Tag anti-collision for fixed and mobile scenario. It can work with moving tags or readers without disturbing the identification process

7.2 Direction for Future Research While RFID provides promising benefits such as inventory visibility and business process automation, some significant challenges need to be overcome before these benefits can be realized. One important issue is how to process and manage RFID data, which is typically in large volume, noisy and unreliable, time-dependent, dynamically changing, and of varying ownership. Finally, RFID systems present a number of inherent vulnerabilities with serious potential security implications. Indeed, given the ability of inexpensively tagging and thus monitoring a large number of items and/or people, RFID raises some serious security and privacy concerns. RFID systems are vulnerable to a broad range of malicious attacks ranging from passive eavesdropping to active interference. RFID privacy and security are stimulating research areas that involve rich

143

CONCLUSIONS AND FUTURE WORK

CHAPTER 7

interplay among many disciplines, such as signal processing, hardware design, supply-chain logistics, privacy rights, and cryptography.

• Suggested research points: 1- Hardware implementation: simulating the operation of the proposed protocols by using FPGA, and checking the timing sequence. 2- Security issues: The designed anti-collision protocols should take the security problem into account, “How to achieve fast identification protocol while preserving the security and privacy in spite of the usage of simple passive tag with the minimum number of used gates (Lightweight Mutual Authentication Protocol) that can be used in low cost passive tag?”. 3- The problems of RFID tag insertion in banknotes: it is interesting to build simple system for checking and counting the banknotes. 4- Moving scenario of readers and tags of the proposed protocols should be more studied and clarified.

144

LIST OF PUBLICATIONS

LIST OF PUBLICATIONS

LIST OF PUBLICATIONS [1]

Usama Sayed Mohamed and Mostafa Salah, “Parallel Binary Tree Splitting Protocol for Tag Anti-collision in RFID Systems”, Proceeding of the 4th IEEE international workshop in Design and Test (IDT09), Riyadh, Saudia Arabia, Nov. 2009.

[2]

Usama Sayed Mohamed and Mostafa Salah,” Fast and Simple Anticollision Protocol Based on Up-Down Counter and One Bit Reader Response”, accepted in: IWRT- International Workshop on RFID Technology-Concepts, Applications, and Challenges, Portugal,2010.

[3]

Usama Sayed Mohamed and Mostafa Salah,” Integrated Reader and Tag Anti-collision Protocol in RFID Systems based on Similar Topology Trees” submitted to the International Journal of Radio Frequency Identification Technology and Applications (IJRFITA).

[4]

Usama Sayed Mohamed and Mostafa Salah,” Fast Anti-collision Protocol for Fixed and moving Scenario in RFID Systems”, Accepted in The 3rd National Information Technology Symposium (NITS 2011), Riyadh, Saudi Arabia, Mar 6, 2011 . Usama Sayed Mohamed and Mostafa Salah,”Fast and Accurate Tag Anti-collision Algorithm for RFID Systems Based on Parallel Binary Splitting Tree Technique” , accepted in The Sixth International Conference on Computer Engineering & Systems (ICCES'2010), Cairo in the period Nov. 30, 2010 to Dec. 2, 2010.

[5]

145

LIST OF REFERENCES

LIST OF REFERENCES

LIST OF REFERENCES [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18]

K.Finkenzeller, RFID Handbook: Fundamentals and Applications, John Wiley and Sons Ltd, Second Edition, New York, 2003. Hussain Al-Mousawi, “Performance and reliability of Radio Frequency Identification (RFID)”, Masters Thesis in Information and Communication Technology, Grimstad, June 2004. D. Shih, P. -L. Sun, D. C. Yen and S. -M. Huang, “Taxonomy and Survey of RFID Anti-collision protocols,” Computer and communications, 2006. Stephen August Weis, “Security and Privacy in Radio-Frequency Identification Devices”, in partial fulfillment of the requirements for the degree of Master at the MASSACHUSETTS Institute of Technology, May 2003. Ching Law, Kayi Lee, Kai-Yeung Siu, “Efficient Memoryless Protocol for Tag Identification”, MIT- Auto-ID Center, Massachusetts Institute of Technology,Cambridge, http://auto-id.mit.edu/ , October 2000. A. Sahoo, S. Iyer and N. Bhandari, “Improving RFID System to Read Tags Efficiently”, Kanwal Rekhi School of Information Technology, Indian Institute of Technology Bombay, June, 2006 Dong-Her Shih a, Po-Ling Sun, David C. Yen, Shi-Ming Huang, “Taxonomy and survey of RFID anti-collision protocols”, www.elsevier.com/locate/comcom ,2006. Hiroyuki Nakamoto, Daisuke Yamazaki, ” A Passive UHF RF Identification CMOS Tag IC Using Ferroelectric RAM in ٠.35-µm Technology” , IEEE Journal of Solid-State Circuits, VOL. 42, NO. 1, JANUARY 2007. Lehpamer, Harvey, “RFID Design Principles”, chapter 5 “Components of the RFID System, “Artech House Publishers, 2007. [10] Pete Sorrells, “Passive RFID Basics”, 2002 Microchip Technology Inc., U.S.A Zhongxiang Wang, Dan Liu, Xiaofang Zhou, Xi Tan, Junyu Wang, Hao Min, Anti-collision Scheme Analysis of RFID System, Auto-ID Labs White Paper WP-HARDWARE-45,2007. “www.autoidcenter.cn”. Neeraj Chaudhry, Dale R. Thompson,.Craig W. Thompson,” RFID Technical Tutorial and Threat Modeling Version 1” , December 8, 2005. Jung-Sik Cho, Jea-Dong Shin, Sung K. Kim, RFID Tag Anti-Collision Protocol: Query Tree with Reversed IDs, Chung-Ang University, Seoul, Korea, Feb. 1720, 2008 ICACT 2008. Ji H. Choi, Dongwook Lee, and Hyuckjae Lee, “Query Tree-Based Reservation for Efficient RFID Tag Anti-Collision”, IEEE COMMUNICATIONS LETTERS, VOL. 11, NO. 1, JANUARY 2007. Jihoon Myung and Wonjun Lee, “An Adaptive Memoryless Tag Anti-Collision Protocol for RFID Networks”, Korea University, Seoul, Korea, Mar 2005. Shiyu Li and Quanyuan Feng, A Novel Anti-collision Algorithm in RFID System, Progress In Electromagnetics Research Symposium, Cambridge, USA, (2008) 207-210. Ziming Guo, and Binjie Hu, A Dynamic Bit Arbitration Anti-Collision Algorithm for RFID System, IEEE International Workshop on Anticounterfeiting, Security, Identification, (2007) 457 – 460. Mamatha Nanjundaiah, Vipin Chaudhary, “Improvement to the Anti collision Protocol Specification for 900 MHz Class 0Radio Frequency Identification Tag” Advanced Information Networking and Applications, 19th International Conference, (2005) 616- 620 vol.2 146

LIST OF REFERENCES

[19] [20] [21] [22]

[23]

[24] [25] [26] [27]

[28]

[29]

[30]

[31]

Jun-Bo Guo, and Zhen-Hua Ding, “ID-binary tree stack anti-collision algorithm for RFID”, Proceedings of the 11th IEEE Symposium on Computers and Communications, Page(s): 207 - 212, 2006, ISCC '06. Sung Hyun Kim, PooGyeon Park, “An Efficient Tree-Based Tag Anti-Collision Protocol for RFID Systems”, IEEE Communications Letters, vol. 11, No. 5, 449-451 (2007). Jihoon Myung, Wonjun Lee, Jaideep Srivastava, “Adaptive binary splitting for efficient RFID tag anticollision,” Communications Letters, IEEE, Vol. 10, Issue 3, pp.144-146, Mar. 2006. Jihoon Myung, Wonjun Lee, Jaideep Srivastava, and Timothy K. Shih, “Tag splitting: adaptive collision arbitration protocols for RFID tag identification, IEEE Transactions on Parallel and Distributed Systems, Volume: 18 , Issue: 6 , Page(s): 763 – 775, 2007. Wei-Chih Chen, Shi-Jinn Horng, and Pingzhi Fan, An Enhanced Anti- collision Algorithm in RFID Based on Counter and Stack, Proceedings of Second International Conference on Systems and Networks Communications (ICSNC 2007), Issue 25-31, (2007) 21 - 21. J. H. Choi, D.W. Lee and H. J. Lee, “Bi-slotted Tree based Anti-collision Protocols for Fast Tag Identification in RFID Systems,” IEEE Communications Letters, Vol. 10, pp. 861-863, Dec. 2006. Okkyeong Bang, J.H.Choi, D.W. Lee ,H.J.Lee, “Efficient Novel Anti-collision Protocols for Passive RFID Tags,” Auto-ID Labs White Paper WPHARDWARE-050 March 2009. www.autoidlabs.org. Chen-Chung Liu, and Yin-Tsung Chan, “An Anti-Collision protocol of RFID Based on Divide and Conquer Algorithm”, The 26th Workshop on Combinatorial Mathematics and Computation Theory, 2009. Maurizio A. Bonuccelli, Francesca Lonetti, and Francesca Martelli “Tree Slotted Aloha: a New Protocol for Tag Identification in RFID Networks”, Proceedings of the 2006 International Symposium on a World of Wireless, Mobile and Multimedia Networks. J. Waldrop, D. W. Engles, and S. E. Sarma, ”Colorwave : A MAC for RFID Reader Networks,” in Proc. of IEEE Conference on Wireless Communication and Networking, Mar. 2003, Page(s):1206 -1210 vol.2. Dong Wang , Jianwei Wang , Yuping ,”A Novel Solution to the Reader Collision Problem in RFID System”, International Conference on Wireless Communications, Networking and Mobile Computing, 2006, WiCOM 2006, pp.1-4, 22-24 Sept. 2006. Jun-Bong Eom, Tae-Jin Lee,’ RFID Reader Anti-collision Algorithm Using a Server and Mobile Readers Based on Conflict-Free Multiple Access’,27th IEEE International Performance Computing and Communications Conference (IEEE IPCCC 2008), Texas December 7-9, 2008. ETSI EN 302 208-1.2 v1.1.1, September 2004. CTAN: http://www.etsi.org.

[32]

S. M. Birari and S. Iyer, “Mitigating the reader collision problem in RFID networks with mobile readers,” in Proc. 13th IEEE International Conference on Networks, pp. 1-6, Nov. 2005.

[33]

S. R. Lee and C. W. Lee, “RFID reader anti-collision algorithm using the channel monitoring mechanism,” Journal of the Institute of Electronics Engineers of Korea, vol. 43, no. 8, pp. 35-46, Aug. 2006. 147

LIST OF REFERENCES

[34]

[35]

[36] [37] [38] [39]

[40] [41]

[42] [43]

[44]

InChan Song, Xiao Fan, and KyungHi Chang ,’Enhanced Pulse Protocol RFID Reader Anti-collision Algorithm using Slot Occupied Probability in Dense Reader Environment’, KSII Transaction on Internet and Information Systems VOL. 2, NO. 6, December 2008. Leian Liu, Dashun Yan, Xiaozheng Lai, Shengli Lai, “A New Kind of RFID Reader Anti-collision Algorithm”, 4th IEEE International Conference on Circuits and Systems for Communications, ICCSC 2008, 26-28 May 2008 Page(s):559 – 563. Jieun Yu and Wonjun Lee, “GENTLE: Reducing Reader Collision in Mobile RFID Networks”, The 4th IEEE International Conference on Mobile Ad-hoc and Sensor Networks, 10-12 Dec. 2008 Page(s):280 – 287 Kwang-il Hwang, Kyung- tae Kim, and Doo-seop Eom, “DiCA: Distributed Tag Access with Collision-Avoidance Among Mobile RFID Readers” EUC Workshops 2006, LNCS. Dai Hongyue, Lai Shengli, Zhu, Hailong "A Multi-Channel MAC Protocol for RFID Reader Networks" WiCom 2007, 21-25 Sept. 2007, page(s): 2093-2096. Gyanendra Prasad Joshi, Kazi Md. Abdulla Mamun, Sung Won Kim, “A Reader Anti-collision MAC Protocol for Dense Reader RFID System” Proceedings of the IEEE 2009 WRI International Conference on Communications and Mobile Computing, Volume 02, Pages: 313-316. Gyanendra Prasad Joshi and Sung Won Kim, “An Efficient MAC Protocol for Throughput Enhancement in Dense RFID System”, the IEEE 4th International Symposium on Wireless Pervasive Computing, 2009. Gyanendra Prasad Joshi and Sung Won Kim, “Survey, Nomenclature and Comparison of Reader Anti-Collision Protocols in RFID”, The Institution of Electronics and Telecommunication Engineers (IETE), Volume: 25, Issue 5, Page: 285-292, 2008. http://www.rfidjournal.com/ Yuan-Hsin Chen, Shi-Jinn Horng, Ray-Shine Run, Jui-Lin Lai, Rong-Jian Chen, Wei-Chih Chen, Yi Pan, and Terano Takao, ” A Novel Anti-Collision Algorithm in RFID Systems for Identifying Passive Tags”, IEEE Transactions on Industrial Informatics, Vol. 6, No. 1, February 2010. Yuan-Cheng Lai and Chih-Chung Lin, “A blocking RFID anti-collision protocol for quick tag identification”, International Conference on Wireless and Optical Communications Networks, WOCN '09, Cairo, May 2009.

148

‫اﻟﻤﻠﺨﺺ اﻟﻌﺮﺑﻲ‬ ‫ﻓﻲ اﻟﺴﻨﻮات اﻷﺧﻴﺮة أﺻﺒﺤﺖ إﺟﺮاءات اﻟﺘﻌﺮف اﻵﻟﻲ ﻋﻠﻰ اﻟﻬﻮﻳﺔ ذات اﻧﺘﺸﺎر واﺳﻊ ﻓ ﻲ اﻟﻜﺜﻴ ﺮ ﻣ ﻦ‬ ‫اﻟﺼﻨﺎﻋﺎت و اﻟﺨﺪﻣﺎت واﻟﻤﺸﺘﺮﻳﺎت وﺷﺮآﺎت اﻟﺘﺼﻨﻴﻊ و ﺧﻄﻮط اﻻﻧﺘﺎج‪.‬‬ ‫ﺗﻌﺪ ﺗﻜﻨﻮﻟﻮﺟﻴﺎ اﻟﺘﻌﺮف ﻋﻠﻲ ﺑﻄﺎﻗﺎت ﺗﺤﺪﻳﺪ اﻟﻬﻮﻳﺔ ﻋﻦ ﻃﺮﻳﻖ ﻣﻮﺟﺎت اﻟﺮادﻳﻮ ﻣﻦ أآﺜﺮ اﻟﺘﻜﻨﻮﻟﻮﺟﻴﺎت‬ ‫ﺗﻄﻮرا و ﺗﻘﺪﻣﺎ و ذات ﻧﻤﻮ ﻣﺴﺘﻤﺮ‪ .‬وﻣﻦ اﻟﻤﺘﻮﻗﻊ أن هﺬﻩ اﻟﺘﻜﻨﻮﻟﻮﺟﻴﺎ ﺳﺘﻘﺪم ﺗﺤ ﺴﻨﺎ آﺒﻴ ﺮا ﻓ ﻲ اﻟﺘ ﺸﻐﻴﻞ‬ ‫اﻵﻟ ﻲ ‪ ،‬وﻣﺮاﻗﺒ ﺔ اﻟﻤﺨ ﺰون ‪ ،‬وﺗﺘﺒ ﻊ اﻟﻤﻨﺘﺠ ﺎت وﻋﻤﻠﻴ ﺎت اﻟ ﺴﺤﺐ ﻓ ﻲ اﻟﻤﺘ ﺎﺟﺮ واﻟﻤ ﺼﺎﻧﻊ واﻟﺘﺠ ﺎرة‬ ‫واﻟﻨﻘ ﻞ واﻹﻣ ﺪاد ‪ ،‬واﻷﻣ ﻦ ‪ ،‬وﻣ ﺎ إﻟ ﻰ ذﻟ ﻚ ‪ .‬ﺣﻴ ﺚ ﻳﻤﻜ ﻦ ان ﺗ ﺘﻢ ﻋﻤﻠﻴ ﺎت اﻟﺤ ﺼﺮ اﻟﻴ ﺎ و اﻟﺘﺤ ﺪﻳﺚ‬ ‫اﻟﻤﺴﺘﻤﺮ ﻟﻘﻮاﻋﺪ اﻟﺒﻴﺎﻧﺎت ﺗﻌﻜﺲ ﺑ ﺼﻮرة أﻓ ﻀﻞ اﻟﻌ ﺎﻟﻢ اﻟﺤﻘﻴﻘ ﻲ‪ .‬ﺣﻴ ﺚ ﻳﻤﻜ ﻦ ﺗ ﺼﻮر أﻧ ﻪ ﻓ ﻲ ﻳ ﻮم ﻣ ﻦ‬ ‫اﻷﻳﺎم ‪ ،‬ﺳﻴﻜﻮن آﻞ ﻣﻨﺘﺞ ﻳﻨﺘﺠﺔ اﻻﻧﺴﺎن ﻣﻤﻴﺰ ﺑﻬﻮﻳﺔ ﻓﺮﻳﺪة‪.‬‬ ‫أﻧﻬﺎ ﺗﻜﻨﻮﻟﻮﺟﻴﺎ ﺗﺤﺪﻳﺪ اﻟﻬﻮﻳﺔ اﻻﻟﻴﺔ اﻟﺘﻲ ﺗ ﺴﺘﺨﺪم اﻟﺘ ﺮددات اﻟﻼﺳ ﻠﻜﻴﺔ ﻻﺳ ﺘﻘﺒﺎل اﻟ ﺮﻗﻢ اﻟﻤﺴﻠ ﺴﻞ اﻟﻔﺮﻳ ﺪ‬ ‫اﻟﻤﻌﺮف ﻟﻬﺬة اﻟﺒﻄﺎﻗﺔ اﻟﻤﺮﺗﺒﻄﺔ ﺑﺎﻟﻤﻨﺘﺠﺎت اﻟﻤﻄﻠﻮب اﻟﺘﻌﺮف ﻋﻠﻴﻬﺎ‪.‬‬ ‫و ﺗﺘﻤﻴﺰ هﺬة اﻟﺘﻜﻨﻮﻟﻮﺟﻴﺎ ﺑﺄﻧﻬﺎ ﺗﺘﻐﻠﺐ ﻋﻠﻰ اﻟﻘﻴﻮد اﻟﻤﻔﺮوﺿﺔ ﻋﻠﻰ ﺗﻄﺒﻴﻘﺎت أﺧﺮى ﻣﺜﻞ اﻟﺘﻌ ﺮف اﻵﻟ ﻲ‬ ‫اﻟﺒ ﺎرآﻮد واﻟﺒﻄﺎﻗ ﺎت اﻟﻤﻤﻐﻨﻄ ﺔ وﺑﻄﺎﻗ ﺎت اﻟ ﺪواﺋﺮ اﻟﻤﺘﻜﺎﻣﻠ ﺔ‪ .‬و ﻟﻬ ﺎ ﻣﺰاﻳ ﺎ اﺧ ﺮي ﻋﻠ ﻰ ﺳ ﺒﻴﻞ اﻟﻤﺜ ﺎل‬ ‫ﺗﺤﺪﻳﺪ اﻟﺴﺮﻋﺔ ‪ ،‬وﺗﺸﻔﻴﺮ اﻟﺒﻴﺎﻧﺎت ‪ ،‬وﻃﻮل اﻟﻌﻤﺮ ‪ ،‬ودون اﻟﺤﺎﺟﺔ ﻟﺘﻮﻓﻴﺮ اﻻﺗﺼﺎل اﻟﻤﺒﺎﺷﺮ ﻋﺒ ﺮ ﺧ ﻂ‬ ‫اﻷﻓﻖ‪.‬‬ ‫وﻳﺘﻜ ﻮن ه ﺬا اﻟﻨﻈ ﺎم ﻣ ﻦ ﻗ ﺎرئ اﻟﺒﻄﺎﻗ ﺎت اﻟﻜﻬﺮوﻣﻐﻨﺎﻃﻴ ﺴﻲ)ذا ﻗ ﺪرات ﺣ ﺴﺎﺑﻴﺔ و ﻣﻨﻄﻘﻴ ﺔ ﻋﺎﻟﻴ ﺔ(‬ ‫واﻟﺒﻄﺎﻗﺎت )ذات ﻗﺪرات ﻣﺤﺪودة ﺟﺪا ﻟﺘﻜﻮن رﺧﻴﺼﺔ اﻟﺜﻤﻦ( و ﺑﺮاﻣﺞ ادارة اﻟﺒﻴﺎﻧﺎت‪.‬‬ ‫وﺗﺘﻜﻮن اﻟﺒﻄﺎﻗ ﺎت ﻣ ﻦ داﺋ ﺮة ﻣﺘﻜﺎﻣﻠ ﺔ ﺳ ﻠﺒﻴﺔ)ﻻ ﺗﺤ ﻮي ﻣ ﺼﺪر ﻗ ﺪرة ذاﺗ ﻲ آﺒﻄﺎرﻳ ﺔ( ﻣﺘ ﺼﻠﺔ ﺑﻬ ﻮاﺋﻲ‬ ‫ﺻﻐﻴﺮ ﻷﺳﺘﻘﺒﺎل اﻟﺒﻴﺎﻧﺎت و اﻟﻄﺎﻗﺔ ﻣﻦ اﻟﻘﺎرئ اﻟﻤﻮﺟﻮد ﻓﻲ ﻣﺠﺎﻟﺔ‪.‬‬ ‫و ﺗﻜ ﻮن ﻋﻤﻠﻴ ﺔ اﻟﺘﻌ ﺮف اﻻﻟ ﻲ ﻋﻠ ﻲ ه ﺬة اﻟﺒﻄﺎﻗ ﺎت ﻋﺒ ﺎرة ﻋ ﻦ ﻗ ﺮاءة اﻟ ﺮﻗﻢ اﻟﻤﺴﻠ ﺴﻞ اﻟﻤﻤﻴ ﺰ ﻟﻬ ﺬة‬ ‫اﻟﺒﻄﺎﻗﺎت و ذﻟﻚ ﺑﺎﺳﺘﻘﺒﺎﻟﻬﺎ ﺑﺎﺳﺘﺨﺪام ﻣﺒﺪاء اﻻﺗﺼﺎل ﻋﻦ ﻃﺮﻳﻖ اﻟﻄﺎﻗﺔ اﻟﻤﺮﺗﺪة ﻣﻦ هﻮاﺋﻲ اﻟﺒﻄﺎﻗﺎت‪.‬‬ ‫وﻣﻊ ذﻟﻚ ‪ ،‬ﻷن اﻻﺗﺼﺎل ﻋﺒﺮ ﻧﻔﺲ اﻟﻘﻨﺎة اﻟﻼﺳﻠﻜﻴﺔ اﻟﻤﺸﺘﺮآﺔ ‪ ،‬ﺗ ﺴﺒﺐ ﻣ ﺸﻜﻠﺔ اﻟﺘ ﺼﺎدم اﻟ ﺬي ﻳﺤ ﺪث‬ ‫ﻓﻲ اﻧﺘﻘﺎل اﻹﺷﺎرات ﻣﻦ اﻟﻘﺎرئ أو اﻟﺒﻄﺎﻗﺎت ‪ ،‬ﻣﻤﺎ ﻳﺆدي اﻟﻰ ﺑﻂء ﺗﺤﺪﻳﺪ اﻟﻬﻮﻳ ﺔ‪ .‬ه ﺬا اﻟﺘ ﺼﺎدم ﻳ ﺆﺧﺮ‬ ‫اﻧﺘﻘﺎل اﻟﺒﻴﺎﻧﺎت وﻳﻔﻘﺪهﺎ ﺟﺪواهﺎ‪.‬‬ ‫اﻟﻬﺪف اﻟﺮﺋﻴﺴﻲ ﻣﻦ هﺬﻩ اﻟﺮﺳﺎﻟﺔ هﻮ دراﺳﺔ ﻣﺸﻜﻠﺔ ﺗﺼﺎدم اﻟﺒﻴﺎﻧﺎت ﻓﻲ ﺑﻴﺌ ﺔ ﻣﺘﺘﻌ ﺪدة اﻟﺒﻄﺎﻗ ﺎت و آ ﺬا‬ ‫ﻋﻨﺪﻣﺎ ﻳﺘﻮاﺟﺪ أآﺜ ﺮ ﻣ ﻦ ﻗ ﺎرئ ﻓ ﻲ ﻧﻔ ﺲ اﻟﻤ ﺪي‪ ،‬و دراﺳ ﺔ اﻟﻄ ﺮق اﻟ ﺴﺎﺑﻘﺔ اﻟﺘ ﻲ ﺗ ﻢ ﻧ ﺸﺮهﺎ و ﻣﺤﺎوﻟ ﺔ‬ ‫ﺗﺤﺪﻳﺪ ﻣﻮاﻃﻦ اﻟﻘﺼﻮر ﺑﻬﺎ و ﻣﻦ ﺛﻢ ﻟﻨ ﺘﻤﻜﻦ ﻣ ﻦ ﺗﻘ ﺪﻳﻢ ﻃ ﺮق ﺟﺪﻳ ﺪة و ﻓﻌﺎﻟ ﺔ ﻟﺘﺤﻘﻴ ﻖ اﻟﺤ ﺪ اﻻدﻧ ﻲ ﻣ ﻦ‬ ‫اﻟﺘﺼﺎدم و أﻣﻜﺎﻧﻴﺔ اﻟﺘﻌﺮف ﻋﻠﻲ آﻞ اﻟﻬﻮﻳﺎت ﻟﻠﺒﻄﺎﻗﺎت ﺑﺄﺳﺮع ﻣ ﺎ ﻳﻤﻜ ﻦ ﺑﺄﻗ ﻞ ﻋ ﺪد ﻣ ﻦ اﻟﺒﻴﺎﻧ ﺎت اﻟﺘ ﻲ‬ ‫ﻧﺤﺘ ﺎج ﻟﺘﺒﺎدﻟﻬ ﺎ ﺑ ﻴﻦ اﻟﺒﻄﺎﻗ ﺎت و اﻟﻘ ﺎرئ‪ .‬ﺣﻴ ﺚ وﺟ ﺪ أن اﻟﻄ ﺮق اﻟﻤﻘﺘﺮﺣ ﺔ ذات ﺳ ﺮﻋﺔ ﻋﺎﻟﻴ ﺔ ﻣﻘﺎرﻧ ﺔ‬ ‫ﺑﻄﺮق أﺧﺮي ﻣﻌﺮوﻓﺔ ﻓﻲ ﻧﻔﺲ اﻟﻈﺮوف‪.‬‬ ‫ﻓﻲ اﻟﻔﺼﻞ اﻷول ﺗﻢ اﺳﺘﻌﺮاض اﻷﺟﺰاء اﻟﻤﺨﺘﻠﻔﺔ ﻟﻠﺮﺳﺎﻟﺔ اﻟﻐﺮض ﻣﻨﻬﺎ وإﻟﻘﺎء اﻟﻀﻮء ﻋﻠ ﻲ آ ﻞ ﺟ ﺰء‬ ‫ﻣﻨﻬﺎ‪.‬‬ ‫ﻓﻲ اﻟﻔﺼﻞ اﻟﺜﺎﻧﻲ ﺗﻢ اﺳﺘﻌﺮاض ﺷﺎﻣﻞ ﻟﻬﺬا اﻟﻨﻈﺎم‪ :‬ﻣﻜﻮﻧﺎﺗﺔ و ﻧﻈﺮﻳﺔ ﻋﻤﻠﺔ اﻻﺳﺎﺳﻴﺔ وﻣﻤﻴﺰاﺗﺔ‬ ‫و أﻣﺜﻠﺔ ﻷﺣﺪث ﺗﻄﺒﻴﻘﺎﺗﻪ‪ .‬ﺛﻢ دراﺳ ﺔ واﺳ ﺘﻌﺮاض ﻣ ﺸﻜﻠﺔ ﺗ ﺼﺎدم اﻟﺒﻴﺎﻧ ﺎت ﻓ ﻲ ه ﺬا اﻟﻨﻈ ﺎم‪ ،‬اﺳ ﺒﺎﺑﺔ و‬ ‫ﻣﺼﺎدرة اﺳﺒﺎب ﺻﻌﻮﺑﺔ ﺣﻞ اﻟﻤ ﺸﻜﻠﺔ ﻋ ﻦ ﻣ ﺸﺎآﻞ اﻻﺳ ﺘﺨﺪام اﻟﻤﺘﻌ ﺪد ﻟﻠﻮﺳ ﺎﺋﻂ‪ ،‬و آ ﺬا اﻟﻌﻮاﻣ ﻞ اﻟﺘ ﻲ‬

‫ﻳﺠﺐ ﺗﻮاﻓﺮهﺎ ﻓﻲ اﻟﺒﺮوﺗﻜﻮﻻت اﻟﺘﻲ ﺗﺴﺘﺨﺪم ﻟﺤﻞ هﺬة اﻟﻤﺸﺎآﻞ‪ ،‬و اﻟﻌﻮاﻣ ﻞ اﻟﺘ ﻲ ﺗﺤ ﺪد آﻔﺎﺋ ﺔ و ﻣﻴ ﺰة‬ ‫ﺑﺮﺗﻮآﻮل ﻋﻦ اﺧﺮ‪.‬‬ ‫ﻓ ﻲ اﻟﻔ ﺼﻞ اﻟﺜﺎﻟ ﺚ ﺗ ﻢ ﺗﻘ ﺪﻳﻢ دراﺳ ﺔ ﻟﻌ ﺪد آﺒﻴ ﺮ ﻣ ﻦ اﻟﺒﺮوﺗﻮآ ﻮﻻت اﻟﻤ ﺴﺘﺨﺪﻣﺔ ﻟﺘﻔ ﺎدي و ﺗﻘﻠﻴ ﻞ ﺗ ﺄﺛﻴﺮ‬ ‫اﻟﺘﺼﺎدم اﻟﺤﺎدث ﺑﻴﻦ ﺑﻴﺎﻧﺎت اﻟﺒﻄﺎﻗﺎت‪.‬‬ ‫ﻓ ﻲ اﻟﻔ ﺼﻞ اﻟﺮاﺑ ﻊ ﺗ ﻢ ﺗﻘ ﺪﻳﻢ دراﺳ ﺔ ﻟﻌ ﺪد آﺒﻴ ﺮ ﻣ ﻦ اﻟﺒﺮوﺗﻮآ ﻮﻻت اﻟﻤ ﺴﺘﺨﺪﻣﺔ ﻟﺘﻔ ﺎدي و ﺗﻘﻠﻴ ﻞ ﺗ ﺄﺛﻴﺮ‬ ‫اﻟﺘﺼﺎدم اﻟﺤﺎدث ﻓﻲ ﺑﻴﺌﺔ ﻣﺘﺘﻌﺪدة ﻗﺮاء اﻟﺒﻄﺎﻗﺎت‪.‬‬ ‫ﻓﻲ اﻟﻔﺼﻞ اﻟﺨﺎﻣﺲ ﺗﻢ ﺗﻘﺪﻳﻢ ﻃﺮﻳﻘﺔ ﺟﺪﻳﺪة ﻓﻌﺎﻟﻪ ﻟﻠﺘﻌﺮف ﻋﻠﻲ ﺑﻄﺎﻗﺎت اﻟﻬﻮﻳﺔ ﺑﺄﻗﻞ ﻋ ﺪد ﻣ ﻦ اﻟﺒﻴﺎﻧ ﺎت‬ ‫اﻟﻤﺘﺒﺎدﻟﺔ ﺑﻴﻦ اﻟﻘﺎرئ و اﻟﺒﻄﺎﻗﺔ‪ .‬ﻓﻲ ﺗﻠﻚ اﻟﻄﺮﻳﻘﺔ ﻳﺘﻢ اﺳﺘﺨﺪام ﻣﺴﺎر ﺟﺪﻳﺪ ﻟﻠﺘﻌﺮف اﻟﺒﻄﺎﻗﺎت‪.‬‬ ‫هﺬة اﻟﻄﺮﻳﻘﺔ اﻟﻤﻘﺘﺮﺣﺔ ﺗﺴﺘﻨﺪ ﻋﻠ ﻲ اﻟﺘﻘ ﺴﻴﻢ )اﻟﺘﻔ ﺮع( اﻟﺜﻨ ﺎﺋﻲ اﻟﻤﺘ ﻮازي ﻟﻠ ﺸﺠﺮة اﻟﺜﻨﺎﺋﻴ ﺔ و ذﻟ ﻚ ﺑﺄﺗﺒ ﺎع‬ ‫ﻣ ﺴﺎر ﺟﺪﻳ ﺪ ﻟﻼﺳ ﺘﺠﻮاب ﺑﺎﻟﺘﻘ ﺪم اﻟﻤﺘ ﻮازي ﺧ ﻼل اﻟ ﺸﺠﺮة اﻟﺜﻨﺎﺋﻴ ﺔ‪ .‬اﻟﺤ ﻮار ﺑ ﻴﻦ اﻟﺒﻄﺎﻗ ﺎت و ﻗ ﺎرئ‬ ‫اﻟﺒﻄﺎﻗﺎت ﻳﺤﺘﺎج أﺳﺘﺠﺎﺑﺔ اﻟﺒﻄﺎﻗﺎت ﺑﺄرﺳﺎل ﺑﺖ واﺣﺪة ﻓﻘﻂ ﻣﺘﺒﻮﻋﺔ ﺑﺄﺟﺎﺑﺔ اﻟﻘ ﺎرئ ﺑﺎرﺳ ﺎل ﺑ ﺖ واﺣ ﺪة‬ ‫)ﺣﻮار ﺑ ﺖ ﻟﺒ ﺖ واﺣ ﺪة(‪ .‬اﻟﻤﺨﻄ ﻂ اﻟﻤﻘﺘ ﺮح ﻻ ﻳﺤﺘ ﺎج أرﺳ ﺎل ﻧﻘﻄ ﺔ اﻻﺳ ﺘﺌﻨﺎف ﻣ ﻦ اوراق اﻟ ﺸﺠﺮة‬ ‫اﻟﺜﻨﺎﺋﻴﺔ ﺑﻌﺪ اﻷﻧﺘﻬﺎء ﻣﻦ اﻟﺘﻌ ﺮف ﻋﻠ ﻲ آ ﻞ ﺑﻄﺎﻗ ﺔ ﻓ ﻲ أﻃ ﺮاف اﻟ ﺸﺠﺮة اﻟﺜﻨﺎﺋﻴ ﺔ‪ .‬ﻟ ﺬا ﻓ ﺎن ﻋ ﺪد اﻟﺒﺘ ﺎت‬ ‫اﻟﻤﺮﺳﻠﺔ ﻳﺴﺎوي ﺿﻌﻒ ﻋﺪد ﻧﻘﺎط اﻟﺸﺠﺮة اﻟﺜﻨﺎﺋﻴﺔ ﻟﻠﺒﻄﺎﻗﺎت اﻟﻤﻮﺟﻮدة ﻓﻲ ﻧﻄﺎق اﻟﻘﺎرئ ﺑﺄﺳ ﺘﺜﻨﺎء ﻋ ﺪد‬ ‫ﻧﻘﺎط اﻃﺮاف اﻟﺸﺠﺮة‪ .‬ﻧﺘﺎﺋﺞ اﻟﻤﺤﺎآﺎة ﺑﺄﺳﺘﺨﺪام اﻟﻜﻤﺒﻴ ﻮﺗﺮ أﻇﻬ ﺮت أن ﻣﺨﻄ ﻂ ﻣﻨ ﻊ اﻟﺘ ﺼﺎدم اﻟﻤﻘﺘ ﺮح‬ ‫ﺳﺮﻳﻊ و ﺑﺴﻴﻂ‪ .‬ﻧﺘﺎﺋﺞ اﻟﻤﺤﺎآﺎة ﺗﺒﻴﻦ أن هﺬة اﻟﺘﻘﻨﻴﺔ اﻟﻤﻘﺘﺮﺣﺔ ﺗﺘﻔ ﻮق ﻋﻠ ﻲ ﻣﻌﻈ ﻢ اﻟﺘﻘﻨﻴ ﺎت اﻟﺤﺪﻳﺜ ﺔ ﻓ ﻲ‬ ‫ﻣﻌﻈﻢ اﻟﺤﺎﻻت‪ .‬و ﺗﻢ ﺗﻘﺪﻳﻢ ﺗﻄﻮﻳﺮ ﻟﻬﺬة اﻟﻔﻜﺮة ﻟﺘﺘﻘﻠﻴﻞ آﻤﻴﺔ اﻟﺒﻴﺎﻧﺎت اﻟﻼزم ﺗﺒﺎدﻟﻬﺎ ﻋﻦ ﻃﺮﻳﻖ اﺳﺘﺨﺪام‬ ‫ﺑﻄﺎﻗﺎت ذات اﻣﻜﺎﻧﻴﺔ ارﺳﺎل و اﺳﺘﻘﺒﺎل ﻓﻲ ﻧﻔﺲ اﻟﻮﻗﺖ )اﺗ ﺼﺎل ﻣ ﺰدوج(‪ ،‬و آ ﺬا ﺗﻄﺒﻴﻘﻬ ﺎ ﻋﻠ ﻲ اﻟﺒﻴﺌ ﺔ‬ ‫اﻟﺘﻲ ﺗﺤﺘﻮي ﻋﻠﻲ أآﺜﺮ ﻣﻦ ﻗﺎرﺋﻲ‪.‬‬ ‫ﻓﻲ اﻟﻔﺼﻞ اﻟﺴﺎدس ﺗﻢ ﺗﻘﺪﻳﻢ ﻃﺮﻳﻘﺔ ﺟﺪﻳﺪة ﺗﻌﺘﻤﺪ ﻋﻠﻲ ﺗﻄﻮﻳﺮ اﻟﻄﺮق اﻟﻤﻌﺘﻤﺪة ﻋﻠﻲ ﻋﺪاد ﺛﻨﺎﺋﻲ‪،‬‬ ‫و ﻋﻤﻞ اﻟﻘﺎرئ و اﺳﺘﺨﺪام اﻟﻘﺎرئ ﻟﻌﺪد ﺑﻴﺖ واﺣﺪة آﺄﺳﺘﺠﺎﺑﺔ ﻳﺘﻢ ارﺳ ﺎﻟﻬﺎ ﻓﻘ ﻂ ﻓ ﻲ ﺣﺎﻟ ﺔ اﻟﺘ ﺼﺎدم او‬ ‫اﻻﻧﺘﻬﺎ ء ﻣﻦ اﻟﺘﻌﺮف ﻋﻠﻲ اﺣﺪ اﻟﺒﻄﺎﻗﺎت ﻋﻨﺪ أﺣﺪي ورﻗﺎت اﻟﺸﺠﺮة اﻟﺜﻨﺎﺋﻴ ﺔ‪ .‬و أﺧﻴ ﺮا ﺗ ﻢ اﻟﺘﻌﺎﻣ ﻞ ﻣ ﻊ‬ ‫ﺣﺮآﺔ آﻼ ﻣﻦ اﻟﺒﻄﺎﻗﺎت و اﻟﻘﺎرئ‪.‬‬ ‫اﻟﻔﺼﻞ اﻟﺴﺎﺑﻊ ﺗﻢ اﺳﺘﻌﺮاض ﻣﺎ ﺗﻢ اﻟﺤﺼﻮل ﻋﻠﻴﻪ ﻣﻦ ﻧﺘ ﺎﺋﺞ وﺧﻼﺻ ﺔ ﻣ ﺎ ﺗ ﻢ ﻋﻤﻠ ﻪ ﻓ ﻲ ه ﺬﻩ اﻟﺮﺳ ﺎﻟﺔ‬ ‫ﻣﻊ اﻗﺘﺮاح ﻣﺎﻳﻤﻜﻦ ﻋﻤﻠﻪ آﺎﻣﺘﺪاد ﻟﻬﺬا اﻟﻌﻤﻞ ‪.‬‬

‫ﻛﻠﯿﺔ اﻟﮭﻨﺪﺳﺔ‬

‫ﻗﺴﻢ اﻟﮭﻨﺪﺳﺔ اﻟﻜﮭﺮﺑﯿﺔ‬

‫ﺑﺮوﺗﻮﻛﻮل ﻣﻨﻊ اﺻﻄﺪام ﺑﯿﺎﻧﺎت اﻟﺒﻄﺎﻗﺎت ﻓﻲ أﻧﻈﻤﺔ ﺗﺤﺪﯾﺪ اﻟﮭﻮﯾﺔ ﻋﻦ ﺑﻌﺪ‬ ‫و اﻟﺘﻲ ﺗﻌﻤﻞ ﻓﻲ ﻧﻄﺎق ﺗﺮددات اﻟﺮادﯾﻮ )‪(RFID‬‬ ‫رﺳﺎﻟﺔ ﻣﺎﺟﺴﺘﯿﺮ‬ ‫ﻣﻘﺪﻣﮫ ﻣﻦ‬ ‫اﻟﻤﮭﻨﺪس ‪ /‬ﻣﺼﻄﻔﻲ ﺻﻼح ﻋﺒﺪاﻟﺤﻔﯿﻆ ﻣﺤﻤﺪ‬ ‫إﻟﻲ ﻗﺴﻢ اﻟﮭﻨﺪﺳﺔ اﻟﻜﮭﺮﺑﯿﺔ –ﻛﻠﯿﺔ اﻟﮭﻨﺪﺳﺔ ‪-‬ﺟﺎﻣﻌﺔ أﺳﯿﻮط‬

‫! !!! ﻟﺠﻨﺔ اﻻﺷﺮاف‪:‬‬ ‫د‪ /‬أﺳﺎﻣﺔ ﺳﯿﺪ ﻣﺤﻤﺪ ﺳﯿﺪ‬ ‫ﻛﻠﯿﺔ اﻟﮭﻨﺪﺳﮫ‪-‬ﺟﺎﻣﻌﺔ اﺳﯿﻮط‬ ‫د‪ /‬ﻋﺎﻣﺮ ﻋﺒﺪاﻟﻔﺘﺎح ﻋﻠﻲ‬ ‫ﻛﻠﯿﺔ اﻟﮭﻨﺪﺳﮫ‪-‬ﺟﺎﻣﻌﺔ اﺳﯿﻮط‬

‫ﻟﺠﻨﺔ اﻟﺤﻜﻢ ﻋﻠﻲ اﻟﺮﺳﺎﻟﺔ ‪:‬‬ ‫ا‪.‬د‪ /‬ﻣﺤﻲ ﻣﺤﻤﺪ ھﺪھﻮد‬ ‫ﻧﺎﺋﺐ رﺋﯿﺲ ﺟﺎﻣﻌﺔ اﻟﻤﻨﻮﻓﯿﺔ‬

‫ا‪.‬د‪/‬ھﺎﻧﻲ ﺳﻠﯿﻢ ﺟﺮﺟﺲ‬ ‫ﻛﻠﯿﺔ اﻟﮭﻨﺪﺳﺔ‪-‬ﺟﺎﻣﻌﺔ اﺳﯿﻮط‬

‫د‪ /‬أﺳﺎﻣﺔ ﺳﯿﺪ ﻣﺤﻤﺪ ﺳﯿﺪ‬ ‫ﻛﻠﯿﺔ اﻟﮭﻨﺪﺳﺔ‪-‬ﺟﺎﻣﻌﺔ اﺳﯿﻮط‬

‫د‪ /‬ﻋﺎﻣﺮ ﻋﺒﺪاﻟﻔﺘﺎح ﻋﻠﻲ‬ ‫ﻛﻠﯿﺔ اﻟﮭﻨﺪﺳﺔ‪-‬ﺟﺎﻣﻌﺔ اﺳﯿﻮط‬