Ramp Metering via Feedback-Aided Iterative Learning Control for Freeway Traffic Systems under Iteration-Varying Boundary Conditions

In this paper, a ramp metering strategy based on feedback-aided iterative learning control (ILC) method is developed for density tracking in macroscopic freeway traffic systems under iteration-varying conditions occurring in the boundary sections. A general state space form of the freeway traffic model is formulated where the boundary traffic variables are explicitly included. To deal with the non-repetitiveness caused by different downstream traffic conditions at different iterations, the classical discrete-time ILC is incorporated with real-time feedback module. Boundedness of output tracking error can be guaranteed under iteration-varying downstream density conditions as the iteration tends to infinity. Different from the existing feedback-feedforward ramp metering strategies, the feedback mechanism proposed in this paper utilizes the observed density error in the section close to the boundary. Simulation results show that learning performance and robustness is guaranteed by the proposed feedback-aided ILC, and better performance is obtained than classical ILC.