Distributed Model-Free Adaptive Predictive Control for Urban Traffic Networks

Abstract

Data-driven control without using mathematical models is a promising research direction for urban traffic control due to the massive amounts of traffic data generated every day. This article proposes a novel distributed model-free adaptive predictive control (D-MFAPC) approach for multiregion urban traffic networks. More specifically, the traffic dynamics of the network regions are first transformed into MFAPC data models, and then, the derived MFAPC data models instead of mathematical traffic models serve as the prediction models in the distributed control design. The formulated control problem is finally solved with an alternating direction method of multipliers (ADMM)-based approach. The simulation results for the traffic network of Linfen, Shanxi, China, show the feasibility and effectiveness of the proposed method.