On VANET Routing Protocol Based On Correlation-ship Analysis

Vehicle ad hoc network(VANET) is a kind of ad hoc network technology established between vehicles and the roadside infrastructure, making vehicles be able to communicate with one another. As a key of the intelligent transportation systems, VANET have attracted extensive attention both in academic and industrys. VANET evolved from traditional mobile ad-hoc networks, but it also has its unique features, which makes the design of routing protocols quite difficult.GPS vehicle trajectory data is an important carrier of vehicle history records; it contains a wealth of vehicle behavior information. As a result, to process vehicle trajectory data with effective mining methods and introduce social network analysis in this field will probably give a band new perspective in the research of VANET.The aim of this thesis is to improve the reliability of data transmission protocol, effectively reduces network bandwidth consumption. Along with correlation analysis of the road traffic, real trajectory data is used to abstracted encounter characteristics between vehicles, based on which routing protocol is enhanced, and a routing protocol on corealtion-ship analysis is proposed. The main points are the following sections:First, current vehicle trajectory data processing methods and techniques are analyzed, based on which a pre-processing algorithm to deal with the raw trajectory data is proposed. The proposed algorithm is used for calculating the distance between vehicles, and the processed data showed better usability. Then, transfering time widow is used to process the trajectory data to abtain the social graph of the veicles. Based on the encounters of the vehicles, an algorithm to build and maintain the corelationships of the cars is proposed. The algorithm is capable of storing the encounters in vehicles' storage and updating the data meanwhile.Secondly, by investigating the flow interaction between road segments with real-time and historical traffic data, the correlation-ship of road segments is analyzed. Combine with the correlation-ship between vehicles, a kind of virtual community is build, all the vehicles in the virtual community are associated with abstract social relations.Then, the probability of vehicles meeting each other in virtual communities can be calculated. Acordding to the probability of vehicles' encounters and their social property, the candidate of next hop node can be decided, based on which, a correlation based routing algorithm is proposed.Finally, a simulation frame combining SUMO and Opnet is adopted to simulate traffic and network. SUMO is able to get road network and the routes of vehicles, the results then imported to Opnet, the senarios will be generated. The framework tend to achieve precise control of vehicle mobility. Experimental simulation on the routing algorithm proposed in this paper is then carried out on the simulation frame work.The experimental results show that: the routing algorithm proposed in the paper can improve the delivery ratio of data packets; meanwhile, on-demand routing is able to reduce the overhead cost of routing algorithm.