A NEW APPROACH TO ESTIMATE FREEWAY OD TRAVEL TIME BASED ON DYNAMIC TRAFFIC STATES ESTIMATION USING FEEDBACK ESTIMATOR

Abstract

A new algorithm was proposed for estimating origin-destination (OD) travel time on a freeway corridor. Traffic conditions were dynamically estimated by Kalman filter combined with the macroscopic traffic flow model. OD travel time was then estimated based on the optimal estimates of traffic states. The authors have been proposing an alternative approach to estimate OD travel times by Neural-Kalman Filter (NKF). Two numerical examples showed that a) the new algorithm gave more accurate estimate of OD travel time, comparing to the existing approach by the NKF, and that b) dense allocation of traffic detectors enabled the traffic states estimation on actual freeway corridor with the appropriate estimation precision. Further developments for applying the proposed algorithm to large freeway networks were finally summarised through the results of two numerical analyses.