Physical Layer Security-Enhancing Transmission ... - IEEE Xplore

3 downloads 0 Views 237KB Size Report
Abstract—Ambient backscatter is a new communication tech- nology that utilizes ambient radio frequency signals of other systems to enable battery-free devices ...
ICWMMN2015 Proceedings

Physical Layer Security-Enhancing Transmission Protocol against Eavesdropping for Ambient Backscatter Communication System Jia You‡ , Gongpu Wang∗ , Zhangdui Zhong† Key Laboratory of Rail Traffic Control and Safety, Beijing Jiaotong University, Beijing, China. School of Computer and Information Technology, Beijing Jiaotong University, Beijing, China. Email: ‡ [email protected], ∗ [email protected], † [email protected]

† ‡ State ∗ †

on the communication distance between the reader and the tag [4]–[7]. To further increase the field coverage and communication range, ambient backscatter is newly proposed [8], [9]. Ambient backscatter utilizes ambient radio frequency (RF) signals, such as television (TV) radio, cellular signals, and Wireless Fidelity (Wi-Fi), to enable battery-free tag to communicate with the reader [8], [9]. The tag was motivated by certain ambient wireless signals, instead of fixed-frequency sine/cosine waves. The key idea of the ambient backscatter can be described as follows [8]: (i) The battery-free devices such as a tag or a sensor can transmit 0 or 1 bit through switching the antenna between reflecting and non-reflecting states, i.e., a change of the tag antenna impedance states; (ii) the transmitter can backscatter information at a much lower data rate than the ambient signals so that the receiver can separate the two signals through averaging. Based on such idea, the authors in [8] devised a prototype that two battery-free devices can communicates via ambient backscatter and demonstrate its practical bit error rate (BER) versus distance. In 2014, further efforts were made, where a communication system, named as Wi-Fi backscatter, was designed to connect the battery-free devices with off-the-shelf Wi-Fi devices [9]. Clearly, ambient backscatter, a new technology, can liberate tags and sensor nodes from maintenance-heavy batteries and enable them to communicate with readers or other devices. It is even expected to yield a next generation of RFID products and generate a new trend of RF-powered computing [10]. However, many theories and hardware realization such as signal processing, physical layer security and bit error rate (BER) performance for the ambient backscatter communication systems are different from those for the existing communication systems, and there exist various open problems that worth further research. In this work, we investigate the security problems for ambient backscatter communication systems consisting of one reader and multiple tags. Due to the broadcasting nature of wireless channels, the transmitted information between the tags and the reader can be overheard by eavesdroppers, which can further violate the legitimate communication process [11]. Traditional way to overcome the security problem is utiliz-

Abstract—Ambient backscatter is a new communication technology that utilizes ambient radio frequency signals of other systems to enable battery-free devices, such as tags and sensors, to communicate with each other. This paper investigates the physical layer security problem for ambient backscatter communication systems with multiple tags/sensors. Specifically, a new protocol is suggested to facilitate communication process and enhance the security of data transmission between the multiple tags and the reader. This protocol can lead to much lower bit error rate (BER) for the uplink from the tags to the reader, and meanwhile maintain almost the same BER for the wiretap link from the tags to the eavesdropper. We also propose detection approaches and derive closed-form detection thresholds for the reader and the eavesdropper. It is shown that increasing number of tags can lower BER at the reader receiver. It is also found that the number of tags and the number of training symbols can be optimized to maximize data rate gap between the reader and the the eavesdropper. Interestingly, it is found that the maximum data rate gap, a measure for secrecy capacity, is not obtained at high signal-to-noise ratio (SNR). Simulation results are then provided to corroborate our proposed studies. Index Terms—Ambient backscatter, wireless sensor network, RF-powered devices, eavesdropper, performance analysis, battery-free tag.

I. I NTRODUCTION Radio Frequency Identification (RFID) systems have attracted increasing attentions from both academic circles and industrial communities over the past two decades [1], [2]. It is forecast by IDTechEx company that the total RFID market, including tags, readers and software/services, will be worth $27.31 billion in 2024, more than three times of current values —- $8.89 billion in 2014 [3]. One key technology for RFID systems is radio backscatter, a type of wireless communication by means of reflection rather than radiation. In a typical RFID system consisting of a reader (or named as an interrogator) and a passive tag (also known as transponder), the reader first generates an electromagnetic wave, and the tag receives and backscatters the wave with modulated information bits to the reader. The traditional backscatter requires that the reader generate a continuous carrier wave which will be received and remodulated by the tag [2]. Therefore, the backscattered wave will suffer from a round-trip path loss, which will impose an limit

43

>ĞŐĂĐLJƌĞĐĞŝǀĞƌ

enable the reader obtain as more information as possible from the tags than the reader. The channels between the RF source and the reader, the k-th tag, and the the eavesdropper are denoted as hr , hk , and he respectively. The channels between the k-th tag and the reader, and the eavesdropper are represented by gk , and fk respectively. We assume that all channels are slow-fading. Denote the signal transmitted by the RF source as s(n)ej2πfs n , where fs represents the carrier frequency of the RF source and s(n) is the complex baseband equivalent signal. The signals received by the k-th tag from the RF source is hk s(n)ej2πfs n . The tag will backscatter the signals to transmit its own binary signals B(n). Suppose that the fading inside the circuit of the k-th tag is ηk . Then the backscattered signal by the k-th tag can be given hk ηk s(n)B(n)ej2πfs n . Accordingly, the reader and the eavesdropper will receive the complex baseband equivalent signal after synchronization

ĂǀĞƐĚƌŽƉƉĞƌ

dĂŐϭ

Ś
(Φ0 + Φkˇ )/2, Φ0 > Φkˇ .

e(n) =he s(n) + we (n), n ∈ [1, N0 ] (18) e(n) =he s(n) + ηk hk fk s(n) + we (n), n ∈ [N0 + 1, NP ] (19) (20) e(n) =he s(n) + ηkˇ hkˇ fkˇ B(n)s(n) + we (n), n ∈ [NP + NS + 1, N ]

Clearly, (Φ0 + Φkˇ )/2 is the detection threshold. Our transmission protocol can be summarized as follows:

respectively. ˇ To detect the information bits B(n) transmitted by the k-th tag, the eavesdropper will calculate the following statistics

1) At the beginning of every slot, each K tags transmit N0 bits to the reader following a predetermined order, and the reader calculates Φk defined in (5). 2) The reader will find kˇ that satisfies (11) and send a ˇ symbol of the second subsolt. sinusoid wave at the k-th ˇ 3) The k-th tag will response the sinusoid wave with a ˇ binary sequence whose decimal value is k. 4) The reader detects the binary sequence, translates it into ˆ and check if kˆ equals k. ˇ a decimal value k, ˆ ˇ ˇ 5) If k = k, the reader sends a sinusoid wave; if kˆ = k, the reader keeps silent. ˇ tag transmits NT symbols if it receive a sinu6) The k-th soid wave, and there is no transmission if no sinusoid wave is received. 7) Repeat the above six steps for the next slot.

Θ0 =

1 Θkˇ = N0 ΘB (q) =

1 Nt

ˇ 0 kN 

(21) |e(n)|2 ,

(22)

|e(n + NP + NS )|2 .

(23)

ˇ n=(k−1)N 0 +1 q∗N t n=(q−1)∗Nt +1

Subsequently, the eavesdropper will decide  0, if |ΘB (q) − Θ0 | < |ΘB (q) − Θkˇ | ˘ B(q) = 1, if |ΘB (q) − Θ0 | > |ΘB (q) − Θkˇ |.

(24)

The BER and achievable data rate for the eavesdropper can be found as  ˆ = B(n) , (25) Pb,E = Pr B(n)

B. BER and Data Rate In most practical scenarios, the probabilities for B(n) = 0 and B(n) = 1 are equiprobable. Thus the bit error rate Pb,R of the suggested transmission protocol can be obtained as 

N0 1  |e(n)|2 , N0 n=1

Ra,E =Rs Q(1 − Pb,E )(1 − Psel )/N.

(26)

The difference between data rates Ra,R and Ra,E is



RD =Ra,R − Ra,E ˆ = B(n) (16) Pb,R = Pr B(n)  =Rs Q(Pb,E − Pb,R )(1 − Psel )/N, (27) ˆ = 0|B(n) = 1 = Pr (B(n) = 1) Pr B(n) which can be considered as a measure for secrecy capacity.  ˆ = 1|B(n) = 0 + Pr (B(n) = 0) Pr B(n) V. S IMULATION R ESULTS  1  1 ˆ ˆ In this section, the performance of our proposed transmis= Pr B(n) = 0|B(n) = 1 + Pr B(n) = 1|B(n) = 0 . 2 2 sion protocol is studied. The BERs Pb,R and Pb,E and the data rate gap RD (27) are chosen as figure of merits. The channels Since the number of transmitted data symbols B(n) is Q, hr , hk , and he are complex Gaussian random variables with the achievable data rate for the reader can be found as same variance due to similar distances from the RF source to the reader, the tags, and the eavesdropper. The channels fk (17) Ra,R = Rs Q(1 − Pb,R )(1 − Psel )/N, and gk are modelled as complex Gaussian variables with unit variances. All the channels are assumed as slow-fading, i.e., where Rs is the data rate of the signal s(n) and Psel denotes they remain unchanged for the current slot until the start of ˇ the probability that k-th tag is successfully selected by the the next slot. The variance of the noise is taken as Nwr = 1. reader for the last subslot transmission of NT symbols. Fig. 4 depicts the BER curves of Pb,R and Pb,E versus SNR. The slot length N is set as 170 and training length N0 is chosen as 10. The number of tags K is chosen as 2, 4, and IV. S IGNAL D ETECTION AT E AVESDROPPER 6 respectively. It can be found that larger K can significantly We assume that the eavesdropper can receive signals from reduce the BER Pb,R for the reader, while it will result in little the RF source, all tags, and the reader. For each slot, the eaves- changes in the BER Pb,E for the eavesdropper. dropper can obtain the broadcasting message from the reader Fig. 5 plots the data rate gap RD (27) versus SNR. For ˇ tag will be chosen for transmission. and find that the k-th comparison, four curves when number of tags K = 2, 4, 6, 8

46

VII. ACKNOWLEDGMENT

0

10

This work is partly supported by the NSFC (Grant No. 61401016), State Key Laboratory of Rail Traffic Control and Safety (Grant No. RCS2014ZT33, RCS2014ZQ003), and the Fundamental Research Funds for the Central Universities (Grant No. 2014JBM154).

−1

10

−2

10

BER

K=2,Pb,R −3

K=2, P

10

b,E

R EFERENCES

K=4,Pb,R K=4, Pb,E

−4

10

[1] Y. H. Chen, S. J. Horng, R. S. Run, J. L. Lai, R. J. Chen, W. C. Chen, Y. Pan, and T. Takao, “A Novel Anti-Collision Algorithm in RFID Systems for Identifying Passive Tags,” IEEE Trans. Ind. Informat., vol. 6, no. 1, pp. 105-121, Feb. 2010. [2] C. Boyer and S. Roy, “Backscatter communication and RFID: coding, energy and MIMO analysis,” IEEE Trans. Commun., vol. 62, no.3, pp. 770-785, Mar. 2014. [3] Idtechex. “RFID Forecasts, Players and Opportunities 2014-2024,” July 2014. [Online]. Available: http://www.idtechex.com/research/reports /rfid-forecasts-players-and-opportunities-2014-2024-000368.asp [4] J. D. Griffin and G. D. Durgin, “Complete link budgets for backscatter radio and RFID systems,” IEEE Antennas Propagat. Mag., vol. 51, no. 2, pp. 11-25, Apr. 2009. [5] J. D. Griffin and G. D. Durgin, “Gains for RF tags using multiple antennas,” IEEE Trans. Antennas Propag., vol. 56, no. 2, pp. 563-570, Feb. 2008. [6] J. D. Griffin and G. D. Durgin, “Multipath fading measurement at 5.8GHz for backscatter tags with multiple antennas,” IEEE Trans. Antennas Propag., vol. 58, no. 11, pp. 3694-3700, Nov. 2010. [7] D. Tse and P. Viswanath, Fundamentals of Wireless Communication. Cambridge University Press, 2005. [8] V. Liu, A. Parks, V. Talla, S. Gollakota, D. Wetherall, and J. R. Smith, “Ambient backscatter: wireless communication out of thin air,” in ACM SIGCOMM, Hong Kong, China, 2013, pp. 1-13. [9] B. Kellogg, A. Parks, S. Gollakota, J. R. Smith, and D. Wetherall, “WiFi backscatter: Internet connectivity for RF-powered devices,” in ACM SIGCOMM, Chicago, USA, 2014, pp. 1-12. [10] S. Gollakota, M. S. Reynolds, J. R. Smith, and D. Wetherall, “The emergence of RF-powered computing,” IEEE Computer, vol. 47, pp. 32-39, Jan. 2014. [11] Q. Chi, H. Yan, C. Zhang, Z. Pang, and L. Xu, “A reconfigurable smart sensor interface for industrial WSN in IoT environment,” IEEE Trans. Trans. Ind. Informat., vol. 10, no. 2, pp. 1417-1425 , May 2014. [12] W. Shen, T. Zhang, F. Barac, and M. Gidlund, “PriorityMAC: A priorityenhanced MAC protocol for critical traffic in industrial wireless sensor and actuator networks,” IEEE Trans. Ind. Informat., vol. 10, no. 1, pp. 824-835, Feb. 2014. [13] F. Gandino, B. Montrucchio, and M. Rebaudengo, “Key management for static wireless sensor networks with node adding,” IEEE Trans. Trans. Ind. Informat., vol. 10, no. 2, pp. 1133-1143, May 2014. [14] M. Cheminod, L. Durante, and A. Valenzano, “Review of security issues in industrial networks,” IEEE Trans. Ind. Informat., vol. 9, no. 1, pp. 277-293, Feb. 2013.

K=6,P

b,R

K=6, Pb,E

−5

10

−6

10

0

5

10

Fig. 4.

15 SNR (dB)

20

25

30

BER versus SNR.

5

10

Data rage gap RD

K=2 K=4 K=6 K=8

4

10

3

10

0

5

Fig. 5.

10

15 SNR (dB)

20

25

30

Data rate gap RD versus SNR.

are plotted. It can be seen from Fig. 5 that the maximum data rate gap RD is not achieved when SNR is high. Instead, at medium SNR, our transmission protocol can obtain maximum data rate gap RD . The reason is that both eavesdropper and the reader can decode well at high SNR and thus the data rate gap is reduced. VI. C ONCLUSION Ambient backscatter is a new wireless communication technology with good research potential and large market value as well as many open problems. In this paper, we investigated the physical layer security problem for ambient backscatter communication systems with one reader and multiple tags (or sensors). A transmission protocol was suggested to facilitate the communication process and enhance the security between the tags and the reader. It was found that there existed a tradeoff between the number of training and the number of tags. It was also found that maximum data rate gap, a measure for secrecy capacity, was obtained at medium SNR, instead of high SNR. It is worth noting that there exist many open problems for ambient backscatter communication systems, such as theoretical BER analysis and parameter optimization, which can be good choices for our future work.

47