Adaptive calibration of dynamic speed-density relations for online network traffic estimation and prediction applications

A dynamic speed-density relation is identified by using a transfer function model. The model recognizes the time-lagged response of speed to density as well as autocorrelated system noise. A framework for the adaptive calibration of dynamic speed-density relations in the context of real-time dynamic traffic assignment-simulation operation is presented. The model and the approach are evaluated on the basis of actual sensor data from the Irvine, California, network. The results indicate that use of the transfer function approach in the context of real-time simulation is preferable to the use of conventional static traffic flow models.