SDN Architecture for Intelligent Vehicular Sensors Networks

Intelligent Vehicular Sensors Networks is a vehicle collaboration system enabling vehicles to collect and share sensors data with each other. This collaboration results in the production of new knowledge that makes sense for both applications and users. To ensure a successful integration of intelligence functions in vehicles a reliable and efficient communication system is required. However, in the actual VANET architecture, it is quite difficult to maintain good performance of data transmission. This paper proposes the usage of topology prediction for routing path computation. Based on the regular collected vehicle status, the proposed SDN-VANET Controller is able to predict the trajectory of each vehicle thank to the Kraus microscopic traffic flow model. With the knowledge of routing path, the SDN-VANET Controller generate the corresponding flow table and distribute commands for the concerned vehicles. The main advantage of the proposed routing scheme resides in the fact that it is less affected by the high topology changes in VANET contrary to traditional routing protocols (OLSR, GPSR) which cannot instantly determine the new change of topology. From the simulation results, the SDN solution performs than the traditional routing protocols.