Self-learning method and system for traffic signal control

Abstract

The invention discloses a self-learning method and system for traffic signal control. The method includes the steps that a virtual scene containing an intersection and a virtual annunciator is built according to the shape and geometric dimensions of an actual intersection and road canalization, virtual vehicle detectors are arranged on upstream entrances and upstream exits of all roads of the intersection, and the assigned number of virtual vehicles of different types are generated on all the roads; queuing states of the roads controlled by all phases are obtained through the virtual vehicle detectors; according to a preset lookup table, optimum phase green light timing corresponding to the current queuing states of the roads controlled by all the phases is obtained; after the optimum phase green light timing of all the phases is sequentially executed through the virtual annunciator, average delay of the virtual vehicles at the intersection in the current period is obtained, and an evaluation value in the lookup table is updated to finish self-learning of traffic signal control this time. Through the method and system, the average delay of vehicles at the intersection in actual traffic is effectively reduced, the traffic capacity of the intersection is improved, and the problem that vehicle delay at the intersection is difficult to obtain is solved.

Description

technical field [0001] The invention relates to the technical fields of road traffic control and intelligent traffic, in particular to a self-learning method and system for traffic signal control. Background technique [0002] Intersection is the bottleneck of urban traffic, and its optimal control is the basis for smooth urban traffic. The existing intersection control models are mainly the Webster delay model of the United Kingdom and the HCM2000 delay model of the American Road Capacity Manual. The above two models are mainly composed of uniform arrival delay and random arrival delay items of traffic flow. The Webster delay model can calculate the optimal cycle time of the signal, but it is not suitable for dealing with high saturation intersection traffic flow control (when the saturation is higher than 0.9); the HCM2000 delay model can better estimate the intersection vehicles under different saturation delay scenarios, but does not provide an optimal cycle duration e...

AI Technical Summary

Problems solved by technology

[0006] The purpose of the present invention is to provide a self-learning method and system for traffic signal control, which effectively reduces the average vehicle delay at intersections in actual traffic, improves the traffic capacity of intersections, and solves the problem that vehicle delays at intersections are difficult to obtain question

Images