Design And Implementation Of VANET Routing Algorithm Based On Vehicular Social Attributes Mining

VANET is a technology for building mobile ad hoc networks among vehicles or between vehicles and roadside infrastructure,and it’s an important part of intelligent transportation system.Vehicle routing technology is the hotspot for the research of VANET.The vehicular trajectories data reflects the moving characteristics of the vehicles,through the analysis of these data can obtain the social attributes of the association among vehicles.Based on these attributes,the routing algorithm can be improved.Based on analysis of the GPS trajectories,the situations of the encounter of different modes and the probability of occurrence in specified area are analyzed.After the community division of the nodes,the social attributes of the vehicles can be obtained,then the opportunities routing algorithm based on these attributes is designed.to improve the performance of existed routing algorithms,the social attributes have been applied to them.The routing algorithms have been implemented on simulator to evaluate the performance of them.The main works are listed as follows.First,the encounter model based on trajectories after processed is proposed,and the relationships between the stability of vehicles is mining.The communities of vehicles based on the clique divided and the node connectivity of community and multilevel community attributes are got.The calculation method of the encounter probability of the nodes within the community is given.The vehicle activity area is partition as grid,the nodes’ probability distribution of grid could be obtained.Secondly,based on the social attributes of nodes,the routing algorithm based on social attributes(SARP)is proposed.Based on the vehicular social attributes,the GPSR-R algorithm is improved as GPSR-R-I.And the improved CBR algorithm(CBRI)is proposed by using the nodes’ probability distribution of grid distribution.Finally,the data mining and routing simulation system has been the designed and implemented,which is consisting of data preprocessing model,social attribute mining model.Based on the ONE simulator,the routing algorithms have been simulated,and the real node mobility model is obtained by the format conversion of GPS trajectory.And the routing module on the simulator is modified to achieve a new routing algorithm.The relevant routing algorithm is implemented on the simulator and the experimental results are analyzed.Experiments show that the social attributes of vehicles can guide the design of routing algorithms.The social attributes also can help GPSR-R protocol and CBR protocol reduce the end-to-end transmission delay and improve the delivery ratio in sparse networks.