A Survey of Data-Driven Identification and Signal Control of Traffic Congestion

Traffic congestion is a prevalent traffic phenomenon in many cities all over the world, which leads to traffic safety, environmental pollution, and some other problems, and limits the sustainable development of urban traffic. For the urban road network, more than 50% of the congestion occurs at intersections and nearby areas. To this end, it is necessary to quickly identify the urban network traffic state and the characteristics of spatiotemporal evolution. Accordingly, reasonable traffic control strategies can be developed to reduce traffic delays and alleviate congestion. In recent years, abundant data, powerful computing capability, and advanced machine learning methods allow us to re-examine the entire process from identifying traffic congestion to developing signal control strategies, and many studies have been conducted on this topic. To capture the state-of-the-art in this topic, this paper conducts a survey of data-driven identification and signal control of traffic congestion. On this basis, this paper discusses the challenges faced by data-driven methods and future research directions.