An improved traffic network model for model-based predictive control

The paper improves a macroscopic urban traffic network model for traffic control by taking some real aspects of traffic into account. Compared with the previous model, the new model mainly considers the delay time of vehicles running from the beginning of link to the tail of the waiting queues in link under different traffic situations. Taking each of the models as a prediction model respectively, the MPC controllers are constructed and investigated. Simulation results show that both the MPC controllers based on the two models are capable of controlling the traffic network effectively, but the modified model is more efficient for using as a prediction model for MPC for urban traffic control.