Research On Clustering Routing Mechanism Of Vehicular Ad Hoc Network

With the development of intelligent transportation systems,Vehicular Ad Hoc Networks(VANET)has become a research hotspot in the field of wireless self-organizing networks in recent years.In this paper,the routing mechanism in VANET is studied.In the urban environment,the network environment is complex,with features such as fast node movement,dramatic topology changes and frequent link breaks.These features pose challenges to the design of routing protocols.In this paper,the urban traffic scenarios are studied in two aspects: networking mode and routing mechanism.The main research contents and contributions are as follows:Firstly,in terms of networking,a networking strategy based on clustering structure is adopted.Considering that the vehicle's motion trajectory can be acquired in real time,the ondemand weighted clustering algorithm is improved.On the one hand,a new cluster head capability parameter is defined by selecting three variables: node connectivity,vehicle movement similarity and link life cycle.Based on this,the cluster head election algorithm and cluster generation algorithm are designed.On the other hand,in order to compensate for the large routing overhead and the inability to maintain node roles in common clustering algorithms,a time prediction mechanism is designed,and on this basis,the cluster is maintained.The simulation results show that on the average number of cluster head changes,the weighted clustering algorithm based on motion trajectory adopted in this paper is 63% lower than the minimum ID algorithm and the highest node degree algorithm.It shows that the cluster structure generated by the algorithm of this paper is more suitable for the vehicle self-organizing network environment.Secondly,in terms of routing,in order to improve communication efficiency and achieve network load balancing,the hybrid clustering routing protocol is improved.On the one hand,in the route query phase,table-driven and on-demand-driven routing mechanisms are adopted.In order to improve the forwarding success rate of data packets,relay cluster head forwarding algorithm is designed in the process of inter-cluster communication.On the other hand,in the routing stage,the Dijkstra algorithm is improved to reduce the link bandwidth utilization,and the partial link overload in the routing process is avoided.Under the same simulation conditions,the simulation results show that the load-balanced cluster based routing protocol(LCRP)proposed in this paper is slightly lower than the DSDV protocol in terms of routing overhead and end-to-end delay,but its packet delivery rate and bandwidth utilization perform well.In the best case,they are better than other protocols by 48% and 20% respectively.Finally,the OMNeT++ and SUMO are used to construct the vehicle self-organizing network simulation platform for algorithm simulation.According to the MWCA clustering algorithm proposed in this paper,the network performance simulation of the average cluster head change times and the cluster life cycle is carried out for different communication range and vehicle numbers,which proves that the MWCA algorithm can improve the stability of the cluster.Under the same simulation conditions,the network performance of the LCRP routing protocol proposed in this paper is compared with the typical AODV,DSDV and CRP routing protocols.It proves that the protocol has a good performance in improving transmission reliability and easing network congestion.