ES-band

Abstract

Urban arterial traffic coordination control catches great attention in the process of smart city construction. To achieve the optimum signal timing, many studies attempt to adjust green splits of a cycle time according to the distance between the road intersections. However, the existing green wave traffic control system usually has a sophisticated calculation, which depends upon the stability of vehicle speed and traffic flow, leading to weak robustness. Therefore, this short paper puts forward ES-Band, that is, a novel approach to control arterial traffic coordination with the help of artificial intelligence. ES-Band introduces the Covariance Matrix Adaptation Evolutionary Strategies (CMA-ES), a scalable alternative to reinforcement learning, into signal timing. Different traffic variables are adopted as parameters for searching the optimal value by CMA-ES. In order to evaluate the feasibility and effectiveness of ES-Band, we import the real traffic flow data of Zhongshan Road in Ningbo, Zhejiang Province, China, into traffic environment simulator for training and carry out a series of experiments. The results have shown that the ES-Band outperforms the traditional methods in terms of a better convergence, lower travel time, and fewer stops.