A Novel MPC with Chance Constraints for Signal Splits Control in Urban Traffic Network

Abstract It has been recognized that model predictive control (MPC) approach can be successfully applied to the signal control of the urban traffic systems. In this research, we mainly focus on dealing with the uncertainty of the inflow to the traffic network based on the MPC method. The stochastic process describing the uncertainty and chance constraints are embedded to the mathematic programming problem to prevent the congestion happening on the arteries. A modified MPC algorithm is also developed to solve the novel problem under studies. The simulation results show that the proposed model is closer to the real situation than the deterministic ones. The novel MPC algorithm strictly keeps the traffic flows below the limit of the arteries. Moreover, from computation point of view, the proposed method requires shorter computation time that may meet the real-time control requirement.