摘要
Few studies of vehicle-road cooperative control focus on both traffic lights and vehicles.In addition,the fairness metric is rarely considered.In this paper,a intersection traffic control method based on deep reinforcement learning(DRL)named TTVC,which can realise traffic light to traffic light,connected and automated vehicle(CAV)to traffic light and CAV to CAV cooperative control under mixed traffic flow.Cooperative control among agents is achieved through state interaction and policy sharing.Meanwhile,a reward function with a fairness factor is used for traffic lights.Simulation of urban mobility(SUMO)is used to set up both single intersection and multiple intersections,conducting comprehensive simulation experiments under different traffic flow pressure and penetration rate.The simulation results show that TTvC is superior to other methods in each metrics,indicating that the method can improve traffic efficiency and ensure fairness.
基金
supported by The National Key Research and Development Program of China[grant number 2023YFB2504702]
The Natural Science Foundation of Chongqing[grant number CSTB2022NSCQ-LZX0025]
The Talent Program of Chongqing[grant number CSTC2024YCJHBGZXM0037]
The Science and Technology Research Program of Chongqing Municipal Education Commission[grant number KJZDM202300602].
