Adaptive-filtering based dynamic OD matrix estimation

This paper proposes one method for the dynamic origin-destination (OD) estimate. Based on the improved Sage-Husa adaptive filtering algorithm which carries on recursive filter's observation data obtained by time-variable noise statistics estimator, estimate and revision observation noise statistical property, one can suppresses the divergence of the filter and, increase the filter precision. The experiment results indicated that this kind of method has the good performance.