Performance analysis of RSU-to-vehicle system in vehicular delay tolerant networks

RSU-to-vehicle communication is an important issue in the vehicular delay tolerant network and attracts more and more attention. Single roadside unit is the basic scene and traffic scheduling becomes the center in current research, but the throughput is to be resolved. We accurately define the sparse and dense scenes and derive the required condition and corresponding throughput. In sparse scene the throughput is proportional to the average number of arriving vehicles, while in dense scene that is independent with the vehicle arrival rate and determined by the radio coverage area and speed distribution. The intermediate scene, which is between sparse and dense scenes, is analysed by a M/G/∞ queue model, and interestingly the throughput expression covers that of sparse and dense scenes. Three heuristic scheduling policies are adopted in the simulation, and the approximate result of LDF scheduling policy to that of theoretical curve validates our analysis.