An unbiased average traffic speed estimation method for intelligent transportation system

This paper presents a new unbiased average traffic speed (UARTS) estimation method for intelligent transportation system using GPS and radar gun equipped probe vehicles. Unlike the traditional probe-vehicle-based average traffic speed (ARTS) estimation methods, in the UARTS estimation the probe vehicles send their position, time, and neighbor vehicles´ speeds (which are detected by the radar gun on the probe vehicles) to the control center (CC) to proceed with the maximum-likelihood (ML) estimation. In this way, the estimation accuracy is not affected by the distribution of the probe vehicles´ speed. In this paper, the UARTS scheme as well as the GPS and radar gun equipped probe vehicles are described in detail. The mean and the estimation errors of the UARTS and the ARTS estimators are derived. Through computer simulations, the accuracy of the UARTS estimator is examined under different biased speed, detection range of the radar gun, and density of the probe vehicles. Finally, a real data experiment is performed to compare the accuracy of the proposed UARTS and the traditional ARTS estimators.