Expressway exit ramp and adjacent intersection dynamic signal control method

Abstract

The invention discloses a dynamic signal control method for an expressway exit ramp and an adjacent intersection. The method comprises the steps of obtaining traffic flow data, extracting thunder-vision fused trajectory data, constructing and matching a trajectory, predicting a vehicle trajectory of the adjacent intersection and generating an optimized intersection signal timing strategy in the next period. The method comprises the following steps: firstly, extracting and fusing high-precision vehicle tracks at the exits of a ground lane and a ramp acquired by a radar and a video through a neural network and a re-recognition algorithm, then predicting a vehicle track at an adjacent intersection of a next phase by adopting a generative adversarial network, and finally extracting macro-micro characteristics of the vehicle track at the intersection. According to the method, the signal timing strategy of the intersection in the next period is generated through online training in a multi-layer Q reinforcement learning network considering motion waves and exit ramp vehicle lane change congestion, signal timing of the intersection adjacent to the expressway exit ramp is optimized, the traffic volume of the intersection is increased, and the congestion problem is relieved.

Description

technical field [0001] The invention relates to the field of traffic signal management and control, in particular to a dynamic signal control method for an expressway exit ramp and an adjacent intersection. Background technique [0002] With the continuous improvement of computer level, the continuous development of artificial intelligence and deep learning technology, and the increasingly sophisticated traffic information perception technology, it is possible to realize adaptive traffic signal control by using the microscopic data of intersections. Urban expressway is a relatively closed system, which is connected with ordinary roads through on-ramps. The exit-ramp and its adjacent intersections are the key points of the entire road system and one of the bottleneck points that lead to expressway congestion. Make full use of the expressway exit ramp and ground lane vehicle data to predict the vehicle status at the intersection and combine the real-time detection information ...

AI Technical Summary

Problems solved by technology

Generally speaking, the existing research has achieved certain effects on the signal control of single-point intersections by using reinforcement learning, but limited by the way of collecting traffic information, few studies have used the micro-traffic information of intersections as the state set input of reinforcement learning, based on Research on traffic signal control of micro-traffic information is still immature

Images