130 CHAPTER 5 CONCLUSION AND SCOPE FOR FUTURE EXTENSIONS 5.1 INTRODUCTION The feasibility of direct and wireless multi-hop V2V communication based on WLAN technologies, and the importance of position based routing in VANETs are discussed in this thesis. HCBGR, a new unicast position based greedy routing approach has been proposed to support novel applications and to improve the overall performance of routing in VANETs. This domain is recent and the oldest contributions are less than fifteen years old and most of the contributions were made over the last five years. There are a number of contributions that produced significant results, but the general feeling is that the subject is not mature still and that a lot of work remains to be done. VANETs will not only provide safety and life saving applications, but also will become a powerful communication tool for their users. The new projects recently set up in Europe and other countries in Asia and North America are evidences of the interest of many actors in VANETs. The investors believe that this technology can play a significant role in attaining the objectives, such as traffic congestion avoidance and preventing numerous accidents. From a point of view research, more work is needed to improve the performance of the overall system.
131 5.2 CONCLUSION The primary purpose of this thesis is to design an efficient routing algorithm in VANET environment to improve the performance of existing position based routing approaches in VANETs. The HCBGR approach has been proposed to overcome the limitations of existing position based routing approach in VANETs. PDGRP, a position based routing protocol of VANET, is taken from the recent literature to compare the performance of HCBGR. In the base paper of PDGRP, the frequent network disconnections caused by high speed mobility of vehicles is identified as a major issue. It decreases packet delivery rate, increases packet delay and increases routing overhead in VANET environment. Limitations of PDGRP: In high traffic roads, considering future two hop neighbours alone is not sufficient for efficient packet forwarding. The possible occurrence of link breakage event is unknown and unpredictable, due to high dynamics of vehicles. It results in broken links which decreases throughput, increases packet drops and routing overhead. The next hop selection is done on prediction of future neighbours and it is not reliable at all situations. It doesn t guarantee the delivery of packet to the destination node due to high dynamics of vehicles. When vehicle density increases, the number of hops taken by a packet to reach the destination node increases. It increases end to end delay. When vehicle density decreases, the unavailability of neighbour nodes to forward packets cause packet drops. One node could select a next forwarder from neighbours in its full transmission range. A neighbour that was in the transmission range at the moment, but at the edge, could already have left this range and choosing this neighbour as the next forwarder results in increased routing overhead.
132 Advantages of using HCBGR Approach: The existing protocols (For example PDGRP) are not suitable for supporting both dense and sparsely connected vehicular networks. The HCBGR Approach supports efficient data transmission in both dense and sparsely connected networks. The idea of greedy forwarding, and carry and forward strategies are adopted in HCBGR, to forward the packet to the intended vehicle. This approach is demonstrated to perform well in both dense and sparse networks. Reliability of message delivery is improved by allowing multiple vehicles to actively propagate the message. HCBGR algorithm is completely distributed, since nodes need to communicate with only direct neighbours located within their transmission range. Extensive evaluation outlines the advantages of HCBGR, especially in case of high mobility of vehicles and frequent topology changes. It is designed to find robust paths to deliver data with high data delivery ratio, low control overhead and low packet delay. A weighted score based strategy is introduced to predict the probability that a link would be broken in a certain time period based on the mobility information of vehicles. Implementation of HCBGR Approach: The implementation of HCBGR Approach using ns-2 simulator, performs fair and accurate comparisons of these techniques with a broad range of network parameters including mobility, distance and density of vehicles. The GPSRP, PDGRP and HCBGR were simulated in ns-2.33 using RMM. The simulation results show that HCBGR approach can effectively improve routing performance and will be able to deal with the challenges of highway scenarios. HCBGR outperforms GPSRP and PDGRP significantly in terms of improving packet delivery ratio, minimizing end to end delay and reducing the routing overhead. The performance HCBGR approach is compared with PDGRP
133 Approach. Comparison of the proposed HCBGR approach with other existing approaches shows that this routing algorithm is considerably better than the other routing algorithms. Summary of simulation results: HCBGR provides better packet delivery ratio with varying vehicle densities. Using HCBGR-GC, PDR is improved to about 4.3% in comparison with PDGRP and by using HCBGR- DC, PDR is improved to about 10.5% in comparison with PDGRP.HCBGR provides better packet delivery ratio with varied vehicle distance from packet carrier node. Using HCBGR-GC, the PDR is improved by about 6% in comparison with PDGRP and by using HCBGR-DC, the PDR is improved to about 11% in comparison with PDGRP.HCBGR provides better packet delivery ratio with varying speeds of vehicles. Using HCBGR-GC, the PDR is improved to about 7.8% in comparison with PDGRP and by using HCBGR- DC, the PDR is improved to about 11.6% in comparison with PDGRP. HCBGR provides reduced end to end delay with varied vehicle distance from packet carrier node. Using HCBGR-GC, the average delay is reduced to about 0.32 seconds in comparison with PDGRP and by using HCBGR-DC, the average delay is dramatically reduced to about 0.55 seconds in comparison with PDGRP.HCBGR provides reduced end to end delay with varying vehicle densities. Using HCBGR-GC, the average delay is reduced to about 0.43 seconds in comparison with PDGRP and by using HCBGR-DC, the average delay is reduced to about 0.59 seconds in comparison with PDGRP.HCBGR provides reduced end to end delay with varying speeds of vehicles. Using HCBGR-GC, the average delay is reduced to about 0.46 seconds in comparison with PDGRP and by using HCBGR-DC, the average delay is reduced to about 0.6 seconds in comparison with PDGRP.
134 HCBGR provides reduced routing overhead compared with varying vehicle densities. HCBGR-GC has an average gain of 1 hop and HCBGR-DC has an average gain of 2 hops compared to PDGRP. HCBGR provides reduced routing overhead for varied vehicle distance from packet carrier node. HCBGR-GC has an average gain of 1 hop and HCBGR-DC has an average gain of 2 hops when compared to PDGRP.HCBGR provides reduced routing overhead compared with varying speed of vehicles. HCBGR- DC has an average gain of 1 hop compared to PDGRP. 5.3 SCOPE FOR FUTURE EXTENSIONS Though this thesis has been successful in overcoming the limitations of existing routing approaches and improved the performance of routing in VANET environment, it does not take other areas of VANET into account. The subject not yet saturated and a lot of future research works can be pursued in the following areas. MAC: It is needed that research and industry community come to agreement about a MAC technology for VANETs being practical. The trend is towards an extension of IEEE 802. 11 called DSRC. Bandwidth: Due to the limited bandwidth of channel, there is a need for some techniques for controlling the amount of data sent to the network. This problem is addressed in congestion control. A key task for the future is to properly specify the communication requirements of VANET applications and to derive the corresponding optimal tuning of parameters of the communication system, taking into account the current channel and traffic situation.
135 Mobility Model: The mobility of VANETs cannot be captured by general mobility models of MANETs and special mobility models by making use of traffic flow theory, should be proposed. So the simulation results could be trustworthy. The performance of a routing protocol in VANETs depends heavily on the mobility model, the driving environment, the vehicular density, and many other facts. Therefore, having a universal routing solution for all VANETs application scenarios or a standard evaluation criterion for routing protocols in VANETs is necessary. Simulation: Since experimental evaluation of VANETs is expensive, simulation technique should be improved. In addition, with those real-world experiments, simulation models can be refined, platforms can be further developed, and IT management issues can be investigated more deeply. To show the impact of VANETs on traffic safety and efficiency via simulations, accident and human behaviour models are required, that is, one must understand how drivers will react based on the additional information provided by VANETs. Security: Security is an important issue for routing in VANETs, because many applications will effect life-or-death decisions and illicit tampering can have devastating consequences. The characteristics of VANETs make the secure routing problem more challenging and novel than it is in other communication networks. Another challenge related to routing is efficient data dissemination and data sharing in VANETs. Additional areas for improvement include the integration of privacy and security mechanisms into routing protocols and the establishment of priority routes for emergency and safety messages.
136 Communication Paradigms: This research is focused on unicast VANET routing protocols, but many of the features required to make VANETs a viable technology rely on the use of multicast communication paradigms. Much research has been focused on multicast, geocast and any cast communication in VANETs. A survey of these routing protocols is an open topic for future work. With the massive number of vehicles on roads today, the potential for vehicular ad hoc networks is vast. Consumer and corporate interests promise bright prospects for this technology. Routing protocols for VANETs have separated themselves from other mobile routing protocols, due to the inherent characteristics of communication on roadways. Operational Tests: Simply introducing VANETs will not automatically and monotonically increase safety and efficiency. To gain a better understanding of real-world VANETs, field operational tests must be conducted all over the world.