The invention discloses a traffic light cooperative control method, device and equipment and a computer readable storage medium, and the method comprises the steps: obtaining a signal lamp state set through obtaining the states of traffic lights in a preset region, and then taking the signal lamp state set as the input of a preset reinforcement learning model, and taking the preset signal lamp switching state set as the output of a preset reinforcement learning model to obtain an action state value function, and finally training the action state value function by adopting a preset training algorithm to obtain a signal lamp control strategy. According to the method, the traffic light state is input into the reinforcement learning model for reinforcement learning, and after multiple times of training is carried out through the training algorithm, better performance can be obtained on traffic network scales of various scales, especially on a large-scale traffic network.
本发明公开了交通灯协作控制方法、装置、设备及计算机可读存储介质,通过获取预设区域内交通灯的状态以得到信号灯状态集合,然后,将所述信号灯状态集合作为预设强化学习模型的输入,并将预设信号灯切换状态集合作为预设强化学习模型的输出,以得到动作状态价值函数,最后,采用预设训练算法对所述动作状态价值函数进行训练以得到信号灯控制策略。本发明将交通灯状态输入强化学习模型进行强化学习,并且再经过训练算法进行多次训练后,并且,可以在多种尺度的交通网络规模,特别是大规模交通网络上获得更好的性能。
Traffic light cooperative control method, device and equipment and computer readable storage medium
交通灯协作控制方法、装置、设备及计算机可读存储介质
2022-03-04
Patent
Electronic Resource
Chinese
Traffic light information display method, electronic equipment and computer readable storage medium
European Patent Office | 2024
|Light control method and device, computer equipment and computer readable storage medium
European Patent Office | 2023
|Traffic garment, light control method thereof and computer readable storage medium
European Patent Office | 2022
|Traffic light intersection vehicle control method, device and equipment and readable storage medium
European Patent Office | 2024
|Traffic light time dynamic control method, device and equipment and readable storage medium
European Patent Office | 2023
|