A Reinforcement Learning-Based Method for Economical Passage at Intersections Controlled by Signal Lights of Connected Vehicles

Abstract

The present invention relates to a network-connected vehicle signal light-controlled intersection economical traffic method based on reinforcement learning, which comprises the following steps: obtaining vehicle and roadside environmental status information; building an ecological driving model for intersection traffic: including building a longitudinal dynamics model of a vehicle and building a road The traffic environment model of the side signal light state; using the deep deterministic policy gradient algorithm to construct the ecological driving strategy for the intersection of pure electric networked vehicles, solve the multi-objective optimization problem including the minimum energy consumption of the vehicle battery and the shortest transit time, and obtain the optimal traffic speed spectrum , so as to obtain the optimal energy consumption level through the intersection. The invention applies the deep reinforcement learning algorithm to the actual vehicle intersection traffic control, and has good generalization performance and optimization effect for the dynamic change scene of the remaining time of the signal light.

Description

technical field [0001] The invention relates to the technical field of smart traffic and intelligent networked vehicle control, in particular to a method for economical crossing traffic controlled by signal lights of networked vehicles based on reinforcement learning. Background technique [0002] The rapid development of urbanization has brought great challenges to the road traffic environment, and the increasingly crowded traffic intersections with signal lights have generated a lot of unnecessary energy consumption. The entry and exit of economical intersections has always been the main research focus of traffic decision-making in urban traffic intersections. From the perspective of vehicles, the emergence of intelligent networked vehicles provides a new solution to improve traffic efficiency and alleviate the phenomenon of parking and waiting at urban light-controlled intersections. In addition to being driven by new energy, smart electric vehicles are based on the low-...

AI Technical Summary

Problems solved by technology

The Actor-Critic architecture can directly output action information and has powerful optimization capabilities. The research on the convenient application of ecological driving strategies is not yet mature

Images