An adaptive Kalman filter based traffic prediction algorithm for urban road network

Abstract

Frequent traffic congestion and gridlocks are causing global economies staggering cost in terms of fuel consumption, time wastage, and public health. To rectify this problem, many advocates combining Information and Communication Technologies (ICT) and traffic engineering concepts for better traffic management. Timely and accurate traffic prediction and management are central to the ICT-based Intelligent Transportation Systems (ITS). In this paper, we presented a traffic prediction model based on Kalman filtering theory, which optimizes the prediction of speed by minimizing the variance between the real-time speed measurement and its estimation. The prediction model predicts the speed across high-level roadway segments using historical and real-time speed measurements (spot speed) reported by the vehicles traveling on the urban road network. The performance evaluation of the proposed prediction model includes a number of case studies. Each case study is conducted with different parametric settings to explain the different characteristic of the model. The results show that provided the spot speed measurements don't fluctuate significantly over the time, the proposed model is capable of predicting traffic with 54% more accuracy.