Estimating methods of link travel times based on data fusion technology

In order to exactly predict link travel times, the advantages and disadvantages of existing prediction methods were analyzed, adaptive Kalman filter algorithm was used, and link travel time estimation models were presented by combining traffic data from probe vehicles and loop detection. Adaptive Kalman filter(AKF) algorithm-based link travel time estimation models were compared with loop detector data-based methods and probe vehicles-based methods under the circumstances of peak hours and traffic accident, the average absolute percentages of the computation error were analyzed. Analysis result indicates that AKF algorithm is an effective method that may fuse the traffic data from different sources, its predictive values are closer to the measured values so its prediction accuracy is higher.