Metro Traffic Modeling and Regulation in Loop Lines Using a Robust Model Predictive Controller to Improve Passenger Satisfaction

Automatic traffic regulation plays an important role in high-frequency metro lines. This paper introduces a discrete event nonlinear traffic model for metro loop lines based on the deviations from nominal departure times. This model considers the transfer of knock-on delay from one train to another one and improves on previous models. The model also includes uncertainty to accommodate variation in passenger demand over the course of a work day. The nonlinear uncertain traffic model is approximated by the linear uncertain model which facilitates the design of a robust predictive controller. The objective of control is increasing the passenger satisfaction by minimizing the deviation from the nominal time schedule and keeping headways close to the nominal headway. The linear robust model predictive controller (RMPC) is employed to compensate the disturbances and to regulate traffic in the presence of operational constraints. YALMIP toolbox was used to design the RMPC. Model verification and simulations were carried out using real data from Tehran metro line 4. The simulation results are presented to show the trueness of the model and effectiveness of the proposed control methodology.