Research Of Clustering Based VANET Cooperative Downloading Approach

Vehicular ad hoc network is not only has good performance on traffic safety and intelligent transportation, but also have a better performance on entertainment applications and video conferencing and other information services. Vehicular ad hoc network is more and more valued by people, and has become a research hotspot.Vehicles traveling on the highway speed, and deployed in the highway on both sides of road side unit (RSU) is relatively sparse, and the communication range of each RSU is limited, so the communication dark area exist. When the vehicle cannot complete the download tasks in a RSU range of communication, it need to wait for the vehicle to enter the next RSU to downloading. Vehicle in the dark area has long waiting time and high communication delay. So we can use the free vehicle on the road to provide cooperative downloading services for the target vehicle, the throughput can be increased and communication latency can be reduced. Aiming at the low throughput and longer delay problem of cooperative downloading which was put forward by current vehicular ad hoc networks, the thesis proposes improvements.Aiming at problem of high channel contention cause by too many vehicles in one cluster, under the traffic is large with existing clustering algorithm. On the basis of the original algorithm, depending on the density of vehicles on the road, the dynamic distance between adjacent vehicles make the cluster maintain an appropriate size. The simulation results show that, the improved algorithm reduces network delay effectively.Vehicles are clustered through the proposed algorithm, providing cooperative downloading service for target vehicle. The approach make the fully use of vehicles on the road and more vehicles join the cooperative downloading. By this way, the user downloading throughput is improved. At the same time, the downloading delay was reduced. The thesis is simulated under the conditions of NS2 by adopting the macroscopic motion model of high speed traffic flow. The simulation results show that, the clustering based opposite direction cooperative downloading improved the throughput effectively in dark area.