DBSMA Approach for Congestion Mitigation in VANETs

Available online at www.sciencedirect.com Available online at www.sciencedirect.com

ScienceDirect ScienceDirect

Procedia Computer 00 (2017) 000–000 Available online atScience www.sciencedirect.com Procedia Computer Science 00 (2017) 000–000

ScienceDirect

www.elsevier.com/locate/procedia www.elsevier.com/locate/procedia

Procedia Computer Science 109C (2017) 42–49

The 8th International Conference on Ambient Systems, Networks and Technologies The 8th International Conference on Ambient (ANT 2017)Systems, Networks and Technologies (ANT 2017)

DBSMA DBSMA Approach Approach for for Congestion Congestion Mitigation Mitigation in in VANETs VANETs E. A. Feukeu,* T. Zuva E. A. Feukeu,* T. Zuva Dept. of ICT, Vaal University of Technology, Private Bag X021-Vanderbijlpark, South Africa Dept. of ICT, Vaal University of Technology, Private Bag X021-Vanderbijlpark, South Africa

Abstract Abstract A high rate of historical accidents on public roads motivated the advent of the Wireless Acces in Vehicular Environment ( A high rate of historical accidents on public motivated advent strategies of the Wireless Acces in to Vehicular WAVE) standard. The main objective of roads WAVE was tothe develop and protocols enable Environment inter-vehicular( WAVE) standard. The main objective of WAVE was to develop strategies and protocols to enable communication in view to reduce accident on public roads. Moreover, a successful message exchange can only beinter-vehicular possible if the communication in view reduce accident on public roads. Moreover, a successful exchange can only be possible if the transmission medium is to collision free. Under the Intelligent Transportation Systemmessage (ITS), the cooperative Awareness messages transmission is collisionatfree. Under Intelligent Systemfor (ITS), the cooperative Awareness messages (CAM) have medium to be transmitted the rate of the 10 Hz as per Transportation standard. To account congestion management, the Distributed (CAM) haveControl to be transmitted at the rate 10 Hz asHowever, per standard. account fordensity, congestion management, Distributed Congestion (DCC) mechanism wasofproposed. underTohigher node the DDC becomes the inefficient and Congestion Control (DCC) mechanism was proposed. However, under higher node density, the DDC becomes inefficient and dramatically contribute to the deterioration of the VANET environment. The present work proposes a Dynamic Broadcast Storm dramaticallyAlgorithm contribute(DBSMA) to the deterioration of be the used VANET environment. The present work proposes Dynamic Broadcast Mitigation which can to combat the broadcast storm problem in aa Vehicular Network Storm (VN). Mitigation Algorithm (DBSMA) which can that be used to combathas thea broadcast storm problemthe in effect a Vehicular Network (VN). Results from several simulations confirmed the DBSMA potentiality to conquer of broadcast storm by Results from several simulations confirmed that the DBSMA has a potentiality to conquer the effect of broadcast storm by offering higher robustness compared to DCC with about 130% improved efficiency. Beside its computational simplicity, the offering higher robustness compared to DCC with about 130% improved efficiency. Beside its computational simplicity, the DBSMA is easy to implement. DBSMA is easy to implement. © 2016 The©Authors. Published Elsevier by B.V. 1877-0509 2017 The Authors.by Published Elsevier B.V. © 2016 The under Authors. Published by Elsevier B.V. Peer-review responsibility of the Conference Peer-review under responsibility of the Conference Program Program Chairs. Chairs. Peer-review under responsibility of the Conference Program Chairs. Keywords: WAVE; DSRC; CBR; VANET; Congestion Keywords: WAVE; DSRC; CBR; VANET; Congestion

1. Introduction 1. Introduction The Wireless Access in Vehicular environment (WAVE) or Dedicated Short Range Communication (DSRC) was 1 The Wireless Vehicular environment (WAVE) or Dedicated Range Communication (DSRC) was merely motivatedAccess by theindesire to reduce risks of road accidents . WAVEShort refers to a set of emerging standards for merely motivated by the desire to reduce risks of road accidents 1. WAVE refers to a set of emerging standards for * Corresponding author. Tel.: +27 835743136. * Corresponding Tel.: +27 835743136. E-mail address:author. [email protected] E-mail address: [email protected] 1877-0509 © 2016 The Authors. Published by Elsevier B.V. 1877-0509 ©under 2016responsibility The Authors. of Published by Elsevier B.V. Chairs. Peer-review the Conference Program Peer-review under responsibility of the Conference Program Chairs.

1877-0509 © 2017 The Authors. Published by Elsevier B.V. Peer-review under responsibility of the Conference Program Chairs. 10.1016/j.procs.2017.05.293

2

E. A. Feukeu et al. / Computer Procedia Computer (2017) 42–49 E. A.Feukeu / Procedia Science 00Science (2017)109C 000–000

43

mobile wireless communications supporting the Intelligent Transportation System (ITS) 2. Under the ITS framework, a vehicle can directly communicate with its peers referred to as Vehicle-to-Vehicle (V2V) or communicate directly with other existing telecommunication infrastructure also known as Vehicle-to-Infrastructure (V2I) communication. The WAVE standard is a combination of two other standards namely the IEEE 802.11p and the IEEE 1609 3. The former is designed to handle all operations related to the Medium Access Control (MAC) and Physical (PHY) layers while the later concerns more the operations related to upper layers. Dedicated Short-Range Communication (DSRC) 4 was initially coined in USA 5 by the Federal Communication Commission (FCC) 6. It was developed to support vehicle-to-vehicle and vehicle-to-infrastructure communications. This standard supports vehicle speeds up to 190 km/h, a data rate of 6 Mbps (up to 27 Mbps) and a nominal transmission range of 300 m (up to 1000 m). Tremendous efforts in term of strategies development and techniques to ensure successful design, implementation and integration of the WAVE technology in modern cars have already been done. However, there are still numerous challenges to overcome in order to ensure effective integration of this technology into modern vehicles. One of the most common challenges hindering the successful implementation of the WAVE standard is the excessive collision of safety messages as a result of Broadcast Storm (BSt). Naturally the BSt occurrence is the resultant of the standard provision which allow each vehicle to relay or broadcast safety messages at the frequency rate of 10 Hz. This is because a WAVE Short Message Protocol (WSMP) dedicated for safety application must be broadcasted every 50 ms during the Control Channel (CCH) time interval. This phenomenon generally arises during peak hour when the road is congested as a result of road work, car accident or crash. When several cars are stuck in a traffic jam, the BSt becomes unavoidable. Several works have been done in the field VN congestion management. Some of them include the Proactive-Ad hoc On Demand Distance Vector (Pro-AODV) approach proposed in 8. A Pro-AODV is a protocol that uses information from the AODV routing table to minimize congestion in VANETs. Authors in 9 presented and contrasted two classes of congestion control reactive (Decentralized Congestion Control) and adaptive (LIMERIC). Both approaches control safety message transmission as a function of channel load (i.e. Channel Busy Ratio, CBR). The work developed in 10 is enhanced in 11 by providing a study of the convergence to propose a distributed algorithm for the adaptation of transmission probabilities which takes into account the safety benefit of packets transmitted on each wireless link. In 12, a Machine Learning Congestion Control (ML-CC) strategy consisting of three units which include congestion detection, data control and congestion control was proposed to address the congestion that may occur at intersections. A distributed congestion control strategy based on congestion game theory in a single channel vehicular communication environment where periodical broadcasts play the dominant role for driving safety awareness and notifications was proposed in 13. Authors in 14 proposed an intelligent multi-hop broadcast protocol which employs a fuzzy logic-based approach to determine relay (forwarding) vehicles. The parameters used in the protocol are tuned using a Q-Learning-based mechanism where rewards are determined by checking the reception status at the selected relay nodes 15. A study performed in 7 demonstrated that when several mobiles in the vicinity are sending Cooperative Awareness Messages (CAM) and Decentralized Environment Notification Message (DENM), the wireless channel for C-ITS is said to be full at 2000 packets/sec given a packet size of 400 bytes transmitted at a rate of 6 Mbps. The highest throughput however is achieved at 1200 packets/Sec. A direct inconvenience and implication of BSt is wastage of resources and channel bandwidth. The work presented in this paper merely focus on the category of CAM type message used for cooperative awareness. A new approach called DBSMA used to mitigate the broadcast storm problem in a Vehicular Network (VN) is proposed. The DBSMA approach is simple and easy to implement. The DBSMA take into consideration a safety following distance based on a 3 second rule as well as a mobile velocity to compute a corresponding safety message broadcast time. The remainder of this paper is organized as follows. A brief review of the WAVE standard is presented in section 2. The proposed DBSMA development is described in section 3. Simulation and analysis are presented in section 4. Finally the conclusion is presented in section 5.

E. Feukeu A. Feukeu et al. / Procedia Computer 109C (2017) 42–49 E. A. / Procedia Computer Science Science 00 (2017) 000–000

44

3

2. Brief Review of WAVE Standard In order to minimise data collision and optimise the vehicular Network (VN) resources management, the medium access sharing strategy adopted for a VN was derived from the previous version of the IEEE 802.11 standard. The previous IEEE 802.11 is based on a MAC that employed a mandatory contention–based channel access function called the Distributed Coordination Function (DCF) and an optional centrally controlled channel access function called the Point Coordinated access Function (PCF) 16. For WAVE type applications, greater emphasis is placed on the DCF side 17-19. The DCF adopted for the WAVE standard makes use of Carried Sensed Multiple Access with Collision Avoidance (CSMA/CA) 20,21 which is a modified version of the Enhanced Distributed Channel Access (EDCA) 18, 19 mechanism originally provided by IEEE 802.11e 21,22 that differentiates traffic types based upon different static MAC parameters values 20. The DCF is the basic access mechanism of the IEEE 802.11 MAC; it achieves automatic medium sharing between compatible stations through the use of CSMA/CA. In term of congestion management, the ITS proposes an adaptive algorithm under the Distributed Congestion Control (DCC) field to combat the congestion problem in a vehicular network. The proposed DCC approach is summarized by the information provided in table 1 23. The DCC adjusts its safety message transmission rate as a function of channel load or Channel Busy Ratio (CBR) to reduce congestion problem in VN. However, the DCC efficiency is limited especially when the CBR becomes greater than 60%. Table 1. DCC Approach State

Channel Load or CBR

Packet or Transmit rate

Relaxed

< 30%

10 Hz

Active 1

30 - 39%

5 Hz

Active 2

40 - 49%

3.33 Hz

Active 3

50 - 60%

2.5 Hz

Restrictive

> 60%

2 Hz

3. DBSMA Development Taking into Consideration the ability of the driver to react and avoid any incoming hazard which is ultimately a function of its perception and reaction time; the perception time refers to the time it takes to see a hazard and for the brain to realise that there is a hazard that require immediate action to be taken. Generally, the perception time can be as long as 1/4 to 1/2 of a second 24. The reaction time is how long it takes to move the foot from the accelerator to the brake pedal once the brain understands there is danger. This reaction time can vary from 1/4 to 3/4 of a second. Taking into consideration the reaction and action time, a safe following distance rule 25 and 26 were derived. Based on the mentioned rules, if everything is in order and the driver is fully awake, a minimum 3 second rule must be observed in any circumstance. For effective reaction and action, 27 recommended a Minimum safety Distance (MinDist) of 200 m at the speed of 120km/h. Considering the safety distance described 27 as well as the 3 second observation rules, a DBSMA can be developed as follows a)-Based on the WAVE standard, the maximum communication range in a V2V or V2I is 1000 m. b)-If we assume that any car parked on the roadside is a potential accident generator, it means, even parked on the roadside, the faulty car will need to keep on transmitting a safety message to alert the incoming car about its static position on time. c)-Based on the illustrative scenario depicted in Fig. 1, a 20% tolerance was considered for a maximum transmission range. Therefore, the Figure will only consider 800m as the Maximum Transmission range (RM).

4

E. A. Feukeu et al. / Computer Procedia Computer (2017) 42–49 E. A.Feukeu / Procedia Science 00Science (2017)109C 000–000

45

Max Transmission Range=1000m

60sec at 120km/h

Car1

Faulty Car

0

800 m

To

To

Fig. 1. Concept illustration

If the faulty vehicle is positioned in the middle of the maximum transmission range as depicted in Fig. 1, if the incoming vehicle is moving toward the faulty car at the Speed (Si) where the Maximum Speed (SM) in South Africa highways is 120 km/h, it will take a delay time (To) to arrive at the centre position where the faulty vehicle is parked. To= RM/Si

(1)

e)-The multiplier factor alfa (α) to accommodate for the speed variation of 0 to S M can be defined as: α= (CCHTime/To) x 100

(2)

f)-Considering the minimum safety distance (Dm) described in [27] as well as the 3 seconds rules, the time required (Tr) to be able to react and stop on time before the incoming car crash into the faulty car left on the roadside can be defined as Tr= Dm/Si

(3)

g)- The dynamic broadcast time (T B) can then be defined as follows TB=Tr x α

(4)

After execution of equation 4, the response of the DBSMA was computed and presented in Fig. 2.

WSMP broadcast interval (Sec)

DBSMA Response

26 24 22 20 18 16 14 12 10 8 6 4 2 0

10

20

30

40

50 60 70 80 Vehicle Speed (Km/h)

90

100 110 120

Fig. 2. DBSMA response

In this figure it can be noticed that at 120 km/h, the vehicle will be able to transmit at least 6 times before the lapse of the safety distance. It can also be noticed that for all speed less than 6 km/h, the vehicle will only be able to transmit every 24 seconds, which is theoretically equivalent to at least two messages broadcast per 800 m of

E. Feukeu A. Feukeu et al. / Procedia Computer 109C (2017) 42–49 E. A. / Procedia Computer Science Science 00 (2017) 000–000

46

5

travelling distance. The response of the DBSMA strategy allows us to reduce the congestion problem by spacing the broadcast delay, especially in the case of traffic jam on the highways. In general, in a VN, when there is any obstacle ahead, the incoming vehicle will be informed before its arrival at the obstacle spot. This pre-notification allow most cars to naturally slow down. This action therefore increases vehicular density which can be translated into a higher number of messages broadcasted. But since the rate of broadcasting is low at the stop position or at the speed less than 6 km/h, the network resource will be rarely exhausted. To evaluate the effectiveness of the DBSMA, the Channel Busy Ratio (CBR) metric concept was introduced. The reason of the CBR is because in VANET, no handshaking is performed and there isn’t any other way to access the channel resources availability and utilisation. Based on the IEEE 802.11p standard, the CBR is very important metric used to assess the channel quality especially under CSMA/CA without handshaking mechanism. From the linear formulation of the CBR proposed by the Intelligent Transportation System (ITS) technical report on Decentralized Congestion Control algorithm performance evaluation of 2014 28, a CBR limit can be evaluated using the following equation. CBRLimit = Nsta x a + b

(5)

where a=0.000375, b=0.5 and Nsta= number of nodes CBRLimit is the maximum portion of global channel resources that is used by all ITS stations in the radio range of each other. The parameters a and b are chosen to support different scenarios. The choice of parameter b is driven by typical small crossing city traffic scenarios which account for hidden and non-line-of-sight condition. Each vehicle will check its CBR based on the number of messages received during CCH duration. In order to implement DDC approach, the parameter a was chosen to be equal to 0.2. To ensure that the transmitted message goes through with low collision risk, a maximum allowable CBR prescribed by the standard to ensure a possibility of successful packet transmission and reception is 85%. 4. Simulation and Analysis Several simulations were performed to assess and evaluate the efficiency of the DBSMA approach. A DBSMA is compared against the DCC proposed by the ITS standard. A system without any congestion control algorithm should broadcast WSMP at the frequency of 10Hz. Table 2. Simulation Parameters Sparse

10 Cars/ lane

Moderate

40 Cars/lane

Congested

80 Cars/lane

Number of lane

8

Transmission range

1 Km

average Mobility

120 Km/h

Simulation period

120 seconds

This system does not take into consideration the velocity of the involved station. A simulation was performed in MATLAB under different density scenario using parameters presented in table 2. To evaluate the efficiency of the DBSMA against the DCC, Fig. 3 to 8 were computed to account for various scenarios. Three specific scenarios of sparse (10 cars per lane times eight lanes), moderate (40 cars per lane times eight lanes) and congested (80 cars per lane times eight lanes) vehicular density were explored. The comparative

6

E. A. Feukeu et al. / Computer Procedia Computer (2017) 42–49 E. A.Feukeu / Procedia Science 00Science (2017)109C 000–000

47

response of a VN system in a sparse environment with about 80 cars in a radius of 1 km are presented in Fig. 3 in term of number of broadcasted messages and in Fig. 4 in term of CBR. Scenario of 4 lanes per side two ways road in 1 km range

2400

Channel Busy Ratio (%)

Nber of Transmitted Message

0.9

1800 1500 1200 900

0.8 0.7 0.6 0.5

600

0.4

300

0.3

0

0.2 0

10

20

30

40

50 60 70 Time (Sec)

80

DBSMA120 Mod 10 C/lane DCC 120 Mod 10 C/lane

1

DBSMA120 Mod 10 C/lane DCC 120 Mod 10 C/lane

2100

Scenario of 4 lanes per side two ways road in 1 km range

1.1

90

100 110 120

0

Fig. 3 .Message response 10 car/lane of DBSMA vs DCC

10

20

30

40

50 60 70 Time (Sec)

80

90

100 110 120

Fig. 4 . CBR response 10 car/lane of DBSMA vs DCC

A general observation of Fig. 3 and 4 shows that at a channel load less than 60%, the response of the DCC is well regulated. This is true when the simulation period is within the first 10 seconds. However, once the CBR becomes greater than 60%, the DCC response becomes uncontrolled. In these figures, while a DBSMA approach performed under the prescribed limit, a DDC algorithm response breaks the 85% line at approximately 85 second. In both figures, the DBSMA offers an improvement of more than 150% in term of number of transmitted messages and CBR. Scenario of 4 lanes per side two ways road in 1 km range

Scenario of 4 lanes per side two ways road in 1 km range 9000

3.5 DCC 120 Mod 40 C/lane DBSMA120 Mod 40 C/lane

7000 6000 5000 4000 3000 2000

2.5 2 1.5 1 0.5

1000 0

DCC 120 Mod 40 C/lane DBSMA120 Mod 40 C/lane

3

Channel Busy Ratio (%)

Nber of Transmitted Message

8000

0

10

20

30

40

50 60 70 Time (Sec)

80

90

100 110 120

Fig. 5: Message response 40 car/lane of DBSMA vs DCC

0

0

10

20

30

40

50 60 70 Time (Sec)

80

90

100 110 120

Fig. 6: CBR response 40 car/lane of DBSMA vs DCC

A comparative analysis of the DBSMA and DCC in a moderate environment is presented in Fig. 5 and 6. Fig. 5 depicts a number of transmitted message responses while Fig. 6 shows the CBR response of both DCC and DBSMA. Unlike in the case of sparse environment where DBSMA algorithm performed within the required limit, the DCC and DBSMA break the CBR limit at 20 and 54 second respectively. Beside the fact that both break the CBR limit, the DBSMA performance is still clearly observed from Fig. 5 and 6 with about 140 % improvement in both message transmission and CBR. The exploration of Fig.7 and 8 further confirm the robustness of the DBSMA over the DCC. These figures show that as the density increases from 40 to 80 cars per lane, the 85% limit line is also reached faster. This can be confirmed from both figures were the DCC and the DBSMA break the limit line around 10 and 28 seconds respectively.

E. Feukeu A. Feukeu et al. / Procedia Computer 109C (2017) 42–49 E. A. / Procedia Computer Science Science 00 (2017) 000–000

48

Scenario of 4 lanes per side two ways road in 1 km range

8000

6000 5000 4000 3000

DBSMA120 Mod 80 C/lane DCC 120 Mod 80 C/lane

2000 1000 0

0

10

20

30

40

50 60 70 Time (Sec)

80

90

100 110 120

Fig. 7: Message response 80 car/lane of DBSMA vs DCC

Channel Busy Ratio (%)

Nber of Transmitted Message

7000

3 2.8 2.6 2.4 2.2 2 1.8 1.6 1.4 1.2 1 0.8 0.6 0.4 0.2 0

7

Scenario of 4 lanes per side two ways road in 1 km range

DBSMA120 Mod 80 C/lane DCC 120 Mod 80 C/lane 0

10

20

30

40

50 60 70 Time (Sec)

80

90

100 110 120

Fig. 8: CBR response 80 car/lane of DBSMA vs DCC

A general exploration of the DBSMA in comparison to a DCC algorithm clearly demonstrates that the DBSMA offers better performance in all scenario conditions in comparison to the DCC algorithm proposed by the standard. It should be noted that the observed DBSMA performance obtained in the current work was for the worse case condition in all scenarios. This is justified by the fact that in all figures only a mobile speed of 120 km/h was accounted for DBSMA. The DBSMA response in Fig. 2 shows that a higher rate of message transmission is achieved at the speed of 120 km/h. This robustness of the DBSMA clearly confirms its potentiality and strength against congestion in a VN. Holistically, it can be confirmed from simulation response that when the number of broadcasted messages go beyond 2000 messages per second as stated in 7, the CBR response goes above 85% and the wireless channel become saturated. 5. Conclusion The work presented in this paper proposed a new approach called Dynamic Broadcast Storm Mitigation Algorithm (DBSMA) which can be used to combat the broadcast storm problem in a Vehicular Network (VN). A DBSMA concept was explained, the development and derivation steps described and presented. The DBSMA efficiency was evaluated against the DCC approach and its outperformance was observed and demonstrated in all scenarios density conditions. The result from several figures confirmed that the DBSMA has a potentiality to conquer the effect of broadcast storm by offering more than 130% improved efficiency against the DCC approach. Another advantage of the DBSMA is that it is simple to compute and easy to implement. Future work will consider computing the DBSMA under variable mobility scenarios as well as evaluating it in term of channel utilisation. Acknowledgments This work was supported by Vaal University of Technology, the Department of Higher Education and Training (DHET) and the National Research Foundation (NRF) of the Republic of South Africa via the Freestanding Fellowship program. References 1. U.S. Department of transportation., 2006. Vehicle Safety Communications Project Final Report. U.S.: Rep. DOT HS 810 591 Nat. Highway Traffic Safety Admin. 2. Carona. D, Serrador. A, .Mar. P, Abreu. R, Ferreira. N, Meireles. T, Matos. J, Lopes. J, 2010. A 802.11p prototype implementation. IEEE Intelligent Vehicles Symposium University of California. 3. Malathi. V, Charle. L & Brown, 2011. research project from the US DOT FHWA grant no. DTFH61-10-H-00001. Virginia: www.ece.virginia.edu/mv/pubs/./MV-ICC2011-presentation.pdf dept.of Electrical & computer Engineering, University of Virginia.

8

E. A. Feukeu et al. / Computer Procedia Computer (2017) 42–49 E. A.Feukeu / Procedia Science 00Science (2017)109C 000–000

49

4. D. Jiang, V. Taliwal, A. Meier, W. Holfelder, and R. Herrtwich, “Design of 5.9 ghz dsrc based vehicular safety communication,” IEEE Wireless Commun, vol. 13, no. 5, pp. 36–43, Oct. 2006. 5. M. Shulman and R. Deering, “Safety communications in the united states,” Ford Motor Company, General Motors Corporation, United States, Tech. Rep. 07-0010, 2006. 6. Federal Communications Commission, FCC 99-305, FCC Report and Order Std., Oct. 1999. 7. C-ITS Platform, Final report January 2016 [Online] accessed 30 August 2016.http://ec.europa.eu/transparency/regexpert/index.cfm?do=groupDetail.groupDetail&groupID=3188 8. T. Kabir, N. Nurain and M. H. Kabir, "Pro-AODV (Proactive AODV): Simple modifications to AODV for proactively minimizing congestion in VANETs," Networking Systems and Security (NSysS), 2015 International Conference on, Dhaka, 2015, pp. 1-6. doi: 10.1109/NSysS.2015.7043521 9. G. Bansal, B. Cheng, A. Rostami, K. Sjoberg, J. B. Kenney and M. Gruteser, "Comparing LIMERIC and DCC approaches for VANET channel congestion control," Wireless Vehicular Communications (WiVeC), 2014 IEEE 6th International Symposium on, Vancouver, BC, 2014, pp. 1-7. doi: 10.1109/WIVEC.2014.6953217 10. L. Zhang and S. Valaee, “Safety context-aware congestion control for vehicular broadcast networks,” in Proc. 15th IEEE SPAWC, 2014, pp. 399–403. 11. L. Zhang; S. Valaee, "Congestion Control for Vehicular Networks With Safety-Awareness," in IEEE/ACM Transactions on Networking , vol.PP, no.99, pp.1-1. doi: 10.1109/TNET.2016.2521365 12. N. Taherkhani; S. Pierre, "Centralized and Localized Data Congestion Control Strategy for Vehicular Ad Hoc Networks Using a Machine Learning Clustering Algorithm," in IEEE Transactions on Intelligent Transportation Systems , vol.PP, no.99, pp.1-11. doi: 10.1109/TITS.2016.2546555 13. C. Chen, Y. Li, Q. Pei and C. Chen, "Avoiding Information Congestion in VANETs: A Congestion Game Approach," Computer and Information Technology (CIT), 2014 IEEE International Conference on, Xi'an, 2014, pp. 105-110. doi: 10.1109/CIT.2014.53 14. C. Wu, Y. Ji, X. Chen, S. Ohzahata and T. Kato, "An Intelligent Broadcast Protocol for VANETs Based on Transfer Learning, " 2015 IEEE 81st Vehicular Technology Conference (VTC Spring), Glasgow, 2015, pp. 1-6. doi: 10.1109/VTCSpring.2015.7145689 15. C. Watkins, Learning from Delayed Rewards, PhD thesis, King’s College, Cambridge, 1989. 16. Jun CHEN, “Mac-Level Relay in Ad hoc Networks Report of Bibliography”, Master2 RES, pp. 7, April, 2006 17. Part 11: Wireless LAN Medium Access Control (MAC) and Physical Layer (PHY) specifications: Amendment 6: Wireless Access i n Vehicular Environments, IEEE Std. 802.11p-2010, Jun. 2010. 18. Miao, L.; Djouani, K..; Barend, J.; Wyk.V and Yskandar, H., “A Survey of IEEE 802.11p MAC Protocol,” Cyber Journals: Multidisciplinary Journals in Science and Technology, Journal of Selected Areas in Telecommunications (JSAT), September Edition, 2011 19. Chong Han; Dianati, M.; Tafazolli, R.; Kernchen, R.; Xuemin Shen, "Analytical Study of the IEEE 802.11p MAC Sublayer in Vehicular Networks," Intelligent Transportation Systems, IEEE Transactions on , vol.13, no.2, pp.873,886, June 2012 20. Chrysostomou, C.; Djouvas, C.; Lambrinos, L., "Dynamically adjusting the min-max contention window for providing quality of service in vehicular networks," Ad Hoc Networking Workshop (Med-Hoc-Net), 2012 The 11th Annual Mediterranean , vol., no., pp.16,23, 19-22 June 2012 21. Senart,A.; Bouroche, M.; Cahill, V.; Weber, S., Vehicular Networks and Applications, Chapter Book, Springer, 2009. 22. Wireless LAN Medium Access Control (MAC) and Physical Layer (PHY) specifications Amendment 8: Medium Access Control (MAC) Quality of Service Enhancements, IEEE Std. 802.11e, Nov. 2005. 23. C-ITS Platform, Final report January 2016 [Online] accessed 30 August 2016.http://ec.europa.eu/transparency/regexpert/index.cfm?do=groupDetail.groupDetail&groupID=3188 24. LABORATORY, T. R. 2007. Stopping distances for cars [Online]. UK: Road Safety Authority. [Accessed 18July 2016]. 25. Safe following distances (Road rules for everyday driving of the State of Queensland 1995–2016) [Online]. https://www.qld.gov.au/transport/safety/rules/road/distances/index.html accessed on 18 July 2016 26. Following Distances and Road Crashes (13 years of online road safety awareness) [Online] https://arrivealive.co.za/Following-Distances-andRoad-Crashes accessed on 18 July 2016 27. Popular traffic and safety guidelines [Online]. http://www.smartmotorist.com/traffic-and-safety-guideline/maintain-a-safe-following-distancethe-3-second-rule.html accessed on 18 July 2016 28. Intelligent Transport Systems (ITS); Cross Layer DCC Management Entity for operation in the ITS G5A and ITS G5B medium; Report on Cross layer DCC algorithms and performance evaluation. DTR/ITS-0020055/ETSI TR 101612 V1.1.1