A measurement of road segment speed with magnetic sensors

Vehicle speed statistics of road segments is very important for route optimization and traffic control in the Intelligent Transportation System (ITS). Many speed measurements such as probe vehicles based are costly on hardware and reuse of existing surveillance infrastructure. And generally, statistical average speed is used, which is based upon Gaussian distribution is not true in traffic systems. To be more practicable, this paper proposes a weighted averaging method, where speed measurements from multiple magnetic sensors installed on a road segment are weighted-averaged over time adaptively for each sensor and across all lanes to obtain speed index for the road segment. Simulation results show that the road segment weighted average speed is better than statistical average speed in reflecting the real traffic information.