2017 17th IEEE International Conference on Communication Technology
A Collision-Predicted TDMA MAC Protocol in Centralized Vehicular Network
Lizhao Lin, Bin-Jie Hu, Zongheng Wei, Chaodong Wu School of Electronic and Information Engineering South China University of Technology Guangzhou, China e-mail: [email protected]
, [email protected]
, [email protected]
, [email protected]
For the above mentioned advantage of TDMA mechanism, there are many papers using TDMA-based MAC protocol to improve the performance of vehicular networks, while the vehicle’s time slot collision problem there is the main reason for the poor performance of each TDMA protocol. In order to reduce the possibility of access collision and merging collision, which both are time slot collisions, VeMAC protocol [4, 5] uses the idea of time slot reuse and divides each frame into two equal parts for vehicles with different traveling direction. But when some vehicles move faster and traffic density ratio becomes higher, the number of collisions in VeMAC protocol would inevitably increase, thus causing the decline of the channel utilization. To tackle these issues,  proposes a beacon message strategy to evaluate the vehicular density on road constantly and dynamically adjust the slot allocation for different directions according to these beacon messages;  proposes an adaptive vehicular MAC protocol that each vehicle make an adjustment of frame partitioning based on the current traffic conditions in opposite directions before it attempts to reserve a time slot. Thus, both these protocols can significantly improve the channel utilization from the aspect of access collision, especially in the unbalanced traffic scenarios. However, the collision rate of these protocols is still high owing to the remained merging collision problem. To further reduce the collision rate,  proposes a near collision free reservation based MAC protocol (CFR) by dividing one frame into three parts for vehicles in different constant speed level. But in fact the vehicle speed is timevarying. To our best knowledge, merging collision still happen in real life. Borrow the basic idea of A Prediction-based TDMA MAC Protocol (PTMAC) , we propose a collisionpredicted TDMA MAC protocol in centralized vehicular network to decline the possibility of collisions and improve the performance of vehicular network. With the centralized control of RSU, the prediction progress is quiet reliable and collision rate decrease obviously. In Section II, we elaborate the system model, TDMA frame structure, node list definition and relay node selection. In section III, we present the collision-predicted scheme and describe the proposed protocol in details. Section IV gives the simulation results and analysis. Section V comes to a conclusion.
Abstract—Collision is the main problem in Vehicular network, which badly affect network throughput and time delay. To tackle this, this paper proposes a Collision-Predicted TDMA MAC (CPTM) protocol in the centralized vehicular network which includes a Road Side Unit (RSU). The main idea is that the RSU first predicts the merging collisions in the upcoming slots, and then adjusts appropriately the time slots allocation for vehicles to prevent the collisions. In order to decrease the possibility of access collision, empty time slots list is provided by RSU for new vehicles in RSU region. What’s more, RSU reassigns time slots for different-direction vehicles according to vehicle density ratio on roads, which can improve the utilization of channels. We evaluate the proposed protocol with extensive simulations. Simulation results show that comparing with the existing MAC protocols, CPTM can achieve higher throughput and lower time delay. Keywords-vehicular network; MAC protocol; collisionspredicted; RSU
With the development of modern vehicle technique, sensor technique and wireless communication technique, vehicular network is considered a new technology which would improve our life greatly. The United States Federal Communication Commission (FCC) has allocated 75-MHz radio spectrum in the 5.9-GHz band for Dedicated Short Range Communications (DSRC) to be exclusively used by Vehicle to Vehicle and Vehicle to Road Side Unit (RSU) communications . Considering the limited channel resources for vehicular networks, many researches have been made on how to use channel resources effectively and how to improve the key system metrics, such as throughput and time delay, for vehicular networks. Based on the existing literature, there are two main types of mechanism used in MAC protocol in vehicular networks: Carrier sense multiple access/collision avoidance (CSMA/CA) mechanism which is based on competition and time division multiple access (TDMA) mechanism which is based on non-competition . Due to the rapid change of vehicular topology, CSMA/CA mechanism has high possibility of message collisions especially when the traffic density is high . In comparison, TDMA mechanism is more efficient, stable and adaptive, since it enables vehicles to obtain the specific time for transmission which reduce the possibility of message collision greatly.
978-1-5090-3944-9/17/$31.00 ©2017 IEEE
direction vehicles. More vehicles on one direction lanes, relatively more time slots are allocated to those lanes. This allocation is made by RSU using the beacon message strategy in . In this way, the slot utilization is improved and the resource allocation is more reasonable.
SYSTEM MODEL AND STRUCTURE DESIGN
A. System Model Four-way road is the most typical road in most cities in China. In this paper, vehicles run on four-way road. We sort those vehicles into two categories according to their running direction: left direction vehicles and right direction vehicles, as shown in Fig. 1. Vehicles traveling east or easterly direction are defined as the right category, the opposites are defined as the left category. RSU is placed on the roadside, each of them responsible for a specific region on road. Each vehicle is equipped with a global positioning system (GPS) receiver which can help accurately determining its position and traveling direction. Synchronization among vehicles is performed using the 1PPS signal provided by any GPS receiver. In light of the limited transmission range of vehicle, we assume that radius of transmission range is R. Setting the vehicle as center, R as its radius, vehicles beyond the circle cannot receive messages from that centralized vehicle. So the time slot can be reused beyond 2R in vehicle ad hoc network . This paper’s research is based on the time slot reuse. As shown in Fig 1, we divide one RSU region into four parts, the length of each part is twice the transmission range of vehicle (R). In case of the vehicle’s net drop, we set some overlap between two subareas. In each subarea, we select one relay node which responsible for submit information to RSU, the submitted information includes all information from other nodes in its subarea. Relay nodes are very important, since time slots are reused in each RSU region, RSU cannot listen to every node’s information in its region without the help of relay nodes .
Slots for left running vehicles
Slots for right running vehicles
Relay nodes slots
Figure 2. TDMA frame structure.
C. Node List Definition In vehicular environment, each vehicle maintains its own vehicular beacon messages, including direction, position, velocity, transmitted power, owned slot, etc. And neighbor nodes obtain these messages by listening other nodes’ broadcasting within a two-hop area, then update its Node Information List. The relay nodes obtain all messages in its region in the same way. Node Information List includes five types of information, they are respectively direction, position, and velocity, transmitted power and owned slot, which is denoted by:
^NodeID : Direction , Position ,Velocity , TransmittedPower , OwnedSlot `
Direction is a Boolean variable, when it takes 0 it represents vehicles in left lane while 1 represents vehicles in Right lane. Position is a floating point variable, representing the coordinate of vehicle. Velocity is a floating point variable, representing the true velocity. Transmitted power is a floating point variable, representing vehicular transmitted power. Owned slot is a floating point variable, when it takes 0 it represents this vehicle haven’t possess its own time slot, otherwise it represents which time slot this vehicle possessed. In the relay node slots, relay nodes transmit information to RSU, after that, RSU have all information of all vehicles. According to the direction and position of each vehicle in these information, RSU calculate the density ratio of lane in each direction, then adjust the slot allocation for each direction. The list of slot allocation is shown as follows. ^NodeID : Direction` RSU obtain slot-maintain condition of each vehicle in two begin-point-region (where vehicle enter to RSU region), then select unused slots in this region. After this process, RSU get the Empty Slots List which can be used for new vehicles.
Figure 1. Vehicle moving scene with RSU.
B. Frame Structure Design In this paper, time is partitioned into frames. The duration of the frame is 100 ms and each frame consists of 100 constant-duration time slots . As shown in Fig. 2, we divide this 100 time slots into four parts, the beginning two parts are vehicle slots, the third part is for relay nodes, and the last part is RSU slots. Each vehicle in the RSU region should select a specific vehicular slot, during which it broadcast its messages. To each relay node, beside the vehicular slots an extra slot should be selected in the third part of the frame which for submitting collected information to RSU. RSU slots are for RSU to broadcast some important messages to all nodes in its region. Considering the uneven distribution of traffic flow in the real road, we divide all vehicular slots into two parts according to density ratio of the number of different
^emptynodeID : nodeIDi "`
After the collision predict process, RSU change some nodes’ time slots and update the Slot List, each vehicle update their own slot after receiving RSU’s Slot List.
^NodeID : TimeslotID`
D. Relay Node Selection Relay node is a bridge for vehicles in the subarea and RSU. RSU choose the relay node on the basis of vehicle’s position, velocity and signal strength. The best relay node has the closest position to the central position of the subarea PEi , has the minimum variance to average velocity 9 , and has the highest signal strength ERSSI (d ) . A position closing to central position ensure that the relay node can obtain messages more directly, as many vehicles in its one-hop region as possible. The minimum variance to average velocity means relay nodes could keep relatively position to many vehicles in its subarea during a frame time. The highest signal strength makes sure messages transmitted to RSU successfully. Relay node selection algorithm is stated as follows: Wi w1 Distance w2 Velocity w3 NRSSI di (1) All the three parameters are normalized. Distance means the normalized distance of vehicle node to central position in each subarea, it calculated as follows: Distance
Pi PEi Pi PEi
1) RSU gets all information of all vehicles in its region from relay nodes. In the order of time slot ID, RSU searches all nodes use this certain time slot. 2) According to these nodes’ position and speed information, RSU calculates whether these nodes will drive into a relative distance which is less than 2R in the next time slot period. The formula used by RSU is stated as follows: (5) Vi V j u T t D 2R Vi and Vj are the speed of node i and node jˈT represents the time slot period, here T equals 100ms. D represents the distance of node i and node j . 3) If some collisions are predicted, RSU changes the time slot for vehicle which has the lower speed. RSU puts all the changing information in the Slot List and broadcasts it in RSU slots. B. Overall Process On the basis of the system model in section II, we elaborate the overall process of CPTM protocol below. 1) A new vehicle coming into a RSU region has to listen some time until receiving the RSU messages. According to RSU messages, the new comer chooses a time slot in the Empty Slot List of its own driving direction. 2) In the next time slot period, all vehicles broadcast their message in their certain time slot. 3) RSU selects relay nodes based on Relay Node Selection method. 4) According to messages transmitted by relay nodes, RSU: a) Estimates the vehicle density on road, and then allocates time slots for road with different direction. b) Updates relay nodes, and then puts new relay nodes’ time slot ID in Slot List. c) Predicts the upcoming collisions and changes time slots for certain node, puts the result of changing in Slot List. 5) RSU broadcasts Empty Slot List and Slot List. 6) All vehicles in the RSU region listens to the RSU messages and updates its own possessed slots.
Velocity means normalized velocity subtraction of nodes to all nodes’ average velocity in this subarea, it calculated as follows: Velocity
Vi V V V
NRSSI(di) means normalized signal strength, it calculated as follows: NRSSI (di )
RSSI (di ) RSSI (d ) max RSSI (d ) min
Pi=(xi+yi) means the position of node i. PEi=(xa,ya) means the central position of the subarea which node i belonging to. Vi is the instantaneous velocity of node i, 9 is the average velocity of vehicles in the subarea which node i belonging to. 9 is a parameter calculate by RSU. RSSI(d) is link state model, di Pi PEi xi xa 2 yi ya 2 , RSU select nodes with minimum Wi as relay nodes in each subarea. III.
COLLISION-PREDICTED TDMA MAC PROTOCOL
A. Collision Predict Strategy Two types of transmission collision on time slots can happen: access collision and merging collision. An access collision happens when two or more nodes within two hops of each other attempt to acquire the same available time slot. On the other hand, a merging collision happens when two or more nodes acquiring the same time slot become members of the same two-hop set due to node mobility . The main reason of merging collisions is the relatively speed of vehicles which possess same time slots. If those two vehicles drive in same direction, the back one has a faster speed, then after some time, the distance of these two vehicles could be less than 2R. In collision predict strategy, RSU uses position and speed information to predict the upcoming collision. Main steps of collision predict strategy are:
Simulations are performed in MATLAB. We make comparison between the proposed CPTM protocol and VeMAC protocol  in order to analyze the performance of them. We define the distance of RSU region is 8R on the four-way road. New vehicles running into the RSU region are subject to Passion distribution. λ is a variable which represents the possibility of vehicles enter RSU region. Considering the process of vehicle distribution in RSU region, all of our simulation are under the steady state of vehicles. In our simulation scenario, an 800m×20m urban way with 4 lanes is created. A vehicle can communicate with all the vehicles within its communication range. Each vehicle moves with a constant speed, the initial speed of each vehicle is random of certain speed range. The simulation parameters
are listed in Table I. With the limitation of vehicular numbers on real road, we set the range of lama from 0.5 to 2.7. TABLE I.
In Fig. 5, the time delay contains accessing time delay and collision time delay. If there is not enough empty time slot for new comers, some vehicles have to wait until empty time slot list contain empty slots for them. In this case, the waiting period called accessing time delay. If some collisions happen, the vehicles have to wait a time slot period and then contend time slots in next time slot period. In this case, time delay happen because many messages cannot be transmitted in this process. If vehicles cannot obtain the time slot in next time slot period, more time is needed for transmission. As shown in Fig. 5, CPTM protocol has lower time delay because of efficient broadcast of empty time slot list and less collisions. With the increase of vehicle density, both of two protocols’ time delay increasing, but CPTM protocol has much lower speed.
Parameter Urban way length lane width Speed mean value Speed range Transmission range of vehicles, R Transmission range of RSU λ range
Value 800m 5m 9.7m/s [2.7,16.7] 100m 800m [0.5,2.7]
The simulation scenario with moving vehicles is shown in Fig. 3, where the red ones represent the vehicles in left direction, blue ones represent the vehicles in right direction.
Figure 3. The simulation scenario with moving vehicles
The simulation results are shown in Fig. 4 to 6, Fig. 4 shows the throughput on different condition of ¬ in CPTM protocol and VeMAC protocol. As we can see, when ¬ less than 1.1, VeMAC and CPTM protocol have the approximately same throughput; and then with the gradually increase of ¬ , throughput of CPTM higher than the throughput of VeMAC. Also, the throughput of CPTM gradually increase until ¬ is 1.5; but the throughput of VeMAC has a summit when ¬ equals 1.7, then has a down trend. The reason of these phenomenon is that there is very little collision happen when the number of vehicles is low; while with the increasing of vehicular number, more collisions happen in VeMAC protocol, in comparison, CPTM protocol still has little collisions with the help of collision-predict strategy.
Figure 5. Time delay in different ¬.
Considering the time slot assignment of RSU, we simulate the throughput of those two protocols in the scenario of different vehicle density ratio in different driving direction. We set ¬ equals three. As shown in the Fig. 6, with the increasing of vehicle density ratio in different driving direction, the throughput of CPTM still remains, while throughput of VeMAC decreases sharply.
Figure 6. Throughput in different density ratio in opposite lanes.
Throughput in different ¬.
From above results we can learn that the proposed CPTM protocol has higher throughput and lower time delay than VeMAC protocol. Furthermore, CPTM protocol shows better performance when the network is uneven, especially when the vehicle density ratio is high, while the performance of VeMAC protocol is affected seriously by vehicle density ratio. V.
M. Hadded, P. Muhlethaler, A. Laouiti, R. Zagrouba, and L.A. Saidane, "TDMA-Based MAC Protocols for Vehicular Ad Hoc Networks: A Survey, Qualitative Analysis, and Open Research Issues", IEEE Communications Surveys & Tutorials, pp. 2461-2492, DOI: 10.1109/COMST.2015.2440374, June 2015.  H.M. Dai, X. Zhang, D.C. Yang, "Collaborative MAC Protocol Based on Vehicle Ad-Hoc Network,"Measuring Technology and Mechatronics Automation (ICMTMA), DOI: 10.11.9/ICMTMA. 2016.42, May 2016.  M.J. Booysen, G. J. V. Rooyen, "The Effect of TDMA Frame Size and the Hidden Node Problem on the Performance of a NeighborAware VANET MAC," Innovative Mobile and Internet Services in Ubiquitous Computing (IMIS), DOI: 10.1109/IMIS.2014.40, July 2014.  H.A. Omar, W. Zhuang and L. Li, "VeMAC: A novel multichannel MAC protocol for vehicular ad hoc networks," 2011 IEEE Conference on Computer Communications Workshops (INFOCOM WKSHPS), Shanghai, 2011, pp. 413-418.  H.A. Omar, W. Zhuang, and L. Li, "A TDMA-Based MAC Protocol for Reliable Broadcast in VANETs," IEEE Transactions on Mobile Computing, 12(9):1724–1736, Sept 2013.  C.D Wu, B.J. Hu, Z.H. Wei, L.Z. Lin," An Adaptive TDMA Scheduling Strategy Based on Beacon Message for Vehicular AdHoc Network," 2017 9th IEEE International Conference on Communication Software and Networks, Guangzhou, 2017, pp.255261.  P. Chen, J. Zheng, Y.Y. Wu, "A-VeMAC: An Adaptive Vehicular MAC Protocol for Vehicular Ad Hoc Networks," 2017 IEEE International Conference on Communications (ICC), DOI:10.1109/ICC.2017.7997358, May 2017.  R. Zou, Z. Liu, L. Zhang, et al. ͆A near collision free reservation based MAC protocol for VANETs ͇ 2013 IEEE Wireless Communications and Networing Conference (WCNC̓13), Shanghai, China, pp.1538-1543, Apr. 2013.  X.X. Jiang and D.H.C. Du, "PTMAC: A Prediction-based TDMA MAC Protocol for Reducing Packet Collision in VANET,"IEEE Transactions on Vehicular Technology, pp. 9209-9223, doi:10.1109/TVT.2016.2519442,2016.  A. Upadhyay, M. Sindhwani, S.K. Arora, "Cluster head selection for CCB-MAC protocol by implementing high priority algorithm in VANET,” Electronic Design (ICED), DOI: 10.1109/ICED.2016. 7804617, Aug. 2016.
In this paper, a collision-predicted TDMA MAC protocol (CPTM) is proposed for centralized vehicular network. There are two kinds of collisions are the main obstacles for a better performance in MAC layer of vehicular network: accessing collisions and merging collisions. The proposed CPTM protocol makes efficient use of RSU to avoid these collisions. With the help of the relay nodes which chosen by RSU, RSU predicts the merging collisions in the upcoming slots, and then adjusts appropriately the time slots allocation for vehicles to prevent the collisions. In order to decrease the possibility of access collision, RSU provides empty time slots list for new vehicles in RSU region. Furthermore, RSU reassigns time slots for different-direction vehicles according to vehicle density ratio on roads, which can improve the utilization of channels, also decrease the possibility of collision to some extent. We evaluate the proposed protocol with extensive simulations. Compared with VeMAC protocol, the results shows that CPTM protocol has better performance than VeMAC protocol. CPTM has lower possibility of collision, higher throughput and lower time delay. Furthermore, this protocol can be better applied to the traffic flow with uneven distribution on real road. ACKNOWLEDGMENT This work was supported by the IOT Key Project of the Ministry of Industry and Information Technology (351), Guangzhou key science and technology project of Industry-Academia-Research collaborative innovation (2014Y2-00218), University-Industry Key Project of Department of Education of Guangdong Province (CGZHZD1102).