The invention relates to a signal lamp intersection interconnected automobile acceleration trajectory planning method based on deep reinforcement learning. The method comprises the following steps: constructing a vehicle dynamics model and a signal lamp control model; when the interconnected automobile enters the control area, judging whether the current automobile can pass through the intersection in the nearest green light interval under the condition of road speed limitation according to the automobile information and the road information, and if so, executing the next step; constructing a deep reinforcement learning framework, defining a Markov process and setting a neural network; training the neural network according to the vehicle dynamics model, the signal lamp control model and a deep reinforcement learning algorithm until a preset training termination condition is reached, completing training, and obtaining an acceleration control strategy; and the acceleration control strategy is sent to the interconnected automobile, and the interconnected automobile executes the acceleration control strategy and passes through the intersection according to a set target. Compared with the prior art, the method has the advantages of good safety, short passing time, wide application range and the like.
本发明涉及一种基于深度强化学习的信号灯路口互联汽车加速度轨迹规划方法,包括以下步骤:构建车辆动力学模型和信号灯控制模型;当互联汽车进入控制区域时,根据车辆信息以及道路信息判断当前车辆是否可以在最近绿灯区间且在道路限速下通过路口,若可以,则执行下一步;构建深度强化学习框架,定义马尔科夫过程并设置神经网络;根据车辆动力学模型、信号灯控制模型和深度强化学习算法对神经网络进行训练,直至达到预设的训练终止条件,完成训练,得到加速度控制策略;将加速度控制策略发送至互联汽车,互联汽车执行加速度控制策略,按既定目标通过路口。与现有技术相比,本发明具有安全性好、通行时间短、适用范围广等优点。
Signal lamp intersection interconnected automobile acceleration trajectory planning method
一种信号灯路口互联汽车加速度轨迹规划方法
2024-01-02
Patent
Elektronische Ressource
Chinesisch
IPC: | G08G Anlagen zur Steuerung, Regelung oder Überwachung des Verkehrs , TRAFFIC CONTROL SYSTEMS |
Europäisches Patentamt | 2023
|Europäisches Patentamt | 2023
|