Abstract
A novel method for traffic data fusion by using D-S theory and association rules is proposed. The main goal is to obtain more accurate and complete speed information by fusing the traffic speed information from the FCD (Floating Cars Data) system and the traffic volume information collected by loop detectors. After mining the association rules from the mass of data, the confidence of each association rule, which converts the loop detector's volume into traffic speed is computed. Then the D-S theory is used to carry out data fusion. The experiment shows the effectiveness and reliability of the proposed method.