A novel traffic signal split approach based on Explicit Model Predictive Control

As a consequence of the rapid growth of vehicles in cities, urban traffic congestion has become more and more serious nowadays. As an effective control approach, Model Predictive Control (MPC) has been deeply studied for transportation management. However, the relatively complex on-line computing hinders MPC from being further applied on transportation system, especially for large scale transportation networks. Based on the refined store-and-forward model, we propose a signal split control approach with Explicit Model Predictive Control (EMPC) in this paper, by which the complexity of online optimization can be significantly reduced. With multi-parametric Quadratic Program (mp-QP), our approach can generate explicit signal control law, and then shifts the former repeated online signal split optimization offline. As a result, the real time signal control becomes an easier task by searching the lookup table corresponding to the transportation state. The simulation experiments demonstrate that our EMPC based approach outperforms the traditional fix-time signal control. Meanwhile, compared with traditional MPC based approach, our approach can decrease the total computation complexity, and then increase the applicability for real transportation system.