Highway Work Zone Dynamic Traffic Control Using Machine Learning

The focus of this paper is on the application of intelligent transportation systems to work zone traffic management on highways. More specifically, to provide real-time routing information to drivers as they enter the work zone, to assure optimal distribution of traffic across available routes. This paper introduces the use of reinforcement learning, to provide optimal diversion control for a freeway-arterial or express/collector corridor affected by work zones. The paper presents the methodology, development, and simulated testing and results of the machine learning agent. The approach focuses on providing effective route recommendations through variable message signs in order to minimize system wide delay and congestion due to construction. A micro simulation tool - PARAMICS has been used to train the agent on a model of the 401 freeway in the Greater Toronto Area (GTA). Obtained results demonstrate the high potential of this work zone traffic control approach

[1]  Ben J. A. Kröse,et al.  Learning from delayed rewards , 1995, Robotics Auton. Syst..

[2]  Richard S. Sutton,et al.  Reinforcement Learning: An Introduction , 1998, IEEE Trans. Neural Networks.

[3]  W. Thomas Miller,et al.  UNH_CMAC Version 2.1 The University of New Hampshire Implementation of the Cerebellar Model Arithmetic Computer - CMAC , 1996 .

[4]  Wang,et al.  Review of road traffic control strategies , 2003, Proceedings of the IEEE.

[5]  John Moody,et al.  Learning rate schedules for faster stochastic gradient search , 1992, Neural Networks for Signal Processing II Proceedings of the 1992 IEEE Workshop.