Dynamic OD estimation model of urban network

In order to study on dynamic traffic demand, which is based on dynamic traffic management and control, the time-varying relation between OD flows and intersection turning counts was analyzed. By introducing intersection turning counts as additional system measurement, a state-space model considering both link counts and intersection turning counts was established, and a new dynamic OD estimation model of urban network was set up, an efficient method based on inequality constrained Kalman filter was proposed to estimate OD matrices dynamically. The proposed model and method were tested by a synthetic network, which was generated by microscopic traffic simulation tool Paramics. The simulation results show that applying the new model, the absolute and relative estimation average errors are reduced by 9% and 20% respectively, and it has better tracking ability to the time-variation of dynamic OD flow, the results are superior to the computational results of traditional model that only considers link counts.