An Urban Traffic Signal Control System Based on Traffic Flow Prediction

How to improve travel efficiency and alleviate traffic congestion has long been a key problem in intelligent transportation systems. Traffic signal control is a basic tool for urban traffic management. Traditionally, the optimization of traffic light schedule and the prediction of traffic flows are studied separately. In this paper, we aim to combine these two techniques together and propose an urban traffic signal control system based on traffic flow prediction. The objective is to minimize the total number of blocked vehicles at all signalized intersections in the road network. Firstly, a new framework of urban traffic control system including both traffic flow forecasting and signal control optimization is proposed. Secondly, an adaptive traffic light scheduling strategy is designed to alleviate congestion. To validate the proposed system, experiments are performed on the real-world traffic data provided by the Aliyun Tianchi platform. The comparison results show that the proposed system and the signal control optimization strategy perform well.