Adaptive road shoulder traffic control with reinforcement learning approach

To reduce traffic congestion on the highway, variable speed limits, flow control, and traffic light are used in the current traffic control system. Other strategies for easing traffic congestion include controlling the number of vehicles entering the highway by setting up traffic lights on the ramp and extending the number of lanes by opening the shoulder. Although opening the road shoulder to digest the traffic congestion seems to be very efficient, the current system only opens the road shoulder at a fixed schedule. In this research, we proposed a Reinforcement Learning Approach for Adaptive Road Shoulder Traffic Control (ARSTC) to dynamically change the opening and closing time of the road shoulder. The proposed ARSTC technique is able to adjust to different traffic situations and make a suitable decision that is different from the traditional static scheduling approach for opening the road shoulder. The proposed technique is simulated in the Simulation of Urban Mobility. The results showed that ARSTC can reduce traffic congestion time by adaptively controlling the hard shoulders’ opening time and the traffic flow. Our proposed technique (ARSTC) is able to provide safer and more efficient driving conditions while using the hard shoulder to ease traffic congestion.