The invention belongs to the technical field of intelligent traffic, and relates to an intelligent network connection automobile cooperative control method and device in a non-signalized intersection scene, and the method comprises the steps: employing a non-signalized intersection vehicle cooperative control network model trained by a GA-MADDPG algorithm; state information of intelligent agents i in a control area is input into the non-signalized intersection vehicle cooperative control network model, the non-signalized intersection vehicle cooperative control network model outputs an optimal scheduling scheme # imgabs0 # of each intelligent agent i, the optimal scheduling scheme tt is input into a safety inspection module, and if a safety risk exists, the intelligent agents i are controlled to be in the control area. If yes, At is corrected, otherwise, a final scheduling scheme # imgabs1 # is output, the final scheduling scheme # imgabs2 # is transmitted to all the intelligent agents i and executed, and the technical problem that an existing machine learning method cannot efficiently and stably achieve vehicle cooperative control under the complex non-signalized intersection scene is effectively solved.
本发明属于智能交通技术领域,涉及一种无信号交叉口场景下智能网联汽车协同控制方法及装置,该方法包括使用GA‑MADDPG算法训练的无信号交叉口车辆协同控制网络模型,将控制区域内的智能体i的状态信息输入至所述无信号交叉口车辆协同控制网络模型,所述无信号交叉口车辆协同控制网络模型输出为各个智能体i的最优调度方案#imgabs0#将所述最优调度方案tt输入到安全检验模块,若存在安全风险,则对At进行修正,否则输出最终调度方案#imgabs1#将所述最终调度方案#imgabs2#传输给各个智能体i并执行,有效解决现有机器学习方法无法高效、稳定地实现复杂无信号交叉口场景下车辆协同控制的技术问题。
Intelligent network connection automobile cooperative control method and device in non-signalized intersection scene
一种无信号交叉口场景下智能网联汽车协同控制方法及装置
2024-07-26
Patent
Electronic Resource
Chinese
Non-signalized intersection intelligent driving cooperative control method and electronic equipment
European Patent Office | 2023
|European Patent Office | 2024
|European Patent Office | 2024
|Signalized intersection cooperative passing strategy for autonomous vehicle
European Patent Office | 2022
|European Patent Office | 2022
|