Robust vehicle count estimation on urban signalized links

The performance of urban road networks can be easily evaluated based on vehicle count on links; however, it is not an easy task to measure it efficiently. This paper introduces a robust filtering technique for vehicle count estimation on links at signalized intersections by applying the H-infinity filter for which the traffic model is described by the phenomenon of traffic shockwaves. The method requires only measurement data of a single fixed sensor per lane. The proposed estimator has been validated via a high-fidelity microscopic traffic simulator (Vissim) on a test network based on real-world data. The performance of the model has been evaluated by generating gradually increasing traffic disturbance. The traditional Kalman filter was set as the benchmark over which the H-infinity filter had an advantage raising together with the increase of uncertainty. The simulation-based analysis demonstrated the efficiency of the robust estimation technique.