A Robust Kalman Filter Based Data Fusion Method for Traffic Speed Estimation

Abstract

Traffic data is very important for Intelligent Transportation Systems (ITS). The traffic data is collected by different types of detectors. However, the data collected by these detectors cannot meet the precision and completeness requirement of ITS. To improve traffic data’s completeness and comprehensiveness, this paper addresses a problem when road sections have no global positioning system (GPS) sampling signals based on floating vehicle systems. A robust Kalman Filter fusion algorithm is proposed to fuse the information of associated road sections, of which traffic information is known, and to obtain the traffic information of the road sections without GPS sampling signals. The method is verified with the real GPS traffic data of Hangzhou city, China, and compared with the standard Kalman Filter algorithm. Finally, experimental results show that the proposed fusion method is effective and can provide more precise and comprehensive traffic information for traffic managers.