A network-wide efficient prediction model for urban traffic and its application

Model predictive control (MPC) is an effective and efficient method to coordinate urban traffic networks. In order to control an urban traffic network through MPC, it is necessary to establish a proper urban traffic network model. In this paper, an existing link model is at first improved to describe the main traffic dynamic behaviors more accurately. Then, a general urban traffic network topology is proposed. By modeling the network elements, a macroscopic model for urban traffic networks can be established. A model predictive controller is established based on this model. Simulations are carried out to evaluate the control effect of the model predictive controller.