The invention relates to the technical field of traffic signal control, in particular to a single-intersection signal control method based on a deep reinforcement learning algorithm, which comprises the following steps of: acquiring vehicle existence characteristics in cells discretized by lanes, and calculating the communication capability used by a reward function according to the characteristics; the accuracy of traffic state features and reward functions is improved, and a noise network is added in a network structure, so that the action exploration capability of the algorithm is improved, and the convergence speed of the algorithm is improved; the network structure is mainly divided into three parts: a main network, a target network and an experience pool.
本发明涉及交通信号控制的技术领域,特别是涉及一种基于深度强化学习算法的单交叉口信号控制方法,其通过采集由车道离散化的单元格中的车辆存在性特征,奖励函数使用的通信能力也由该特征计算得到,提高了交通状态特征和奖励函数的准确性,并且在网络结构中加入噪声网络,提高算法的动作探索能力,从而提高了算法的收敛速度;网络结构主要分为三个部分:主网络、目标网路和经验池。
Single-intersection signal control method based on deep reinforcement learning algorithm
一种基于深度强化学习算法的单交叉口信号控制方法
2023-04-28
Patent
Electronic Resource
Chinese
Multi-intersection traffic signal control method based on deep reinforcement learning
European Patent Office | 2024
|Value-based deep reinforcement learning for adaptive isolated intersection signal control
IET | 2018
|Regional boundary main intersection signal control method based on deep reinforcement learning
European Patent Office | 2021
|Multi-intersection traffic signal control method based on deep reinforcement learning
European Patent Office | 2023
|