The invention belongs to the technical field of automobiles, and particularly relates to an intelligent automobile rapid landing method based on deep reinforcement learning decision under a roundabout. The method comprises the following steps: step 1, designing a decision based on deep reinforcement learning in a roundabout scene; 2, designing a decision based on a rule algorithm in a roundabout scene; and step 3, designing a decision based on off-line reinforcement learning. The system comprises a vehicle-mounted sensor module, a sensor driving module, a map module, a positioning module, a sensing module, a navigation path module, a fusion decision module and a control module. According to the method, the adaptation to the complex interaction scene of the roundabout can be realized, the safety decision in the training process of the deep reinforcement learning algorithm is ensured, and meanwhile, the rapid landing of the reinforcement learning algorithm can be realized.
本发明属于汽车技术领域,具体的说是一种环岛下基于深度强化学习决策的智能汽车快速落地方法。方法包括以下步骤:步骤一、设计环岛场景下基于深度强化学习的决策;步骤二、设计环岛场景下基于规则算法的决策;步骤三、设计基于离线强化学习的决策。系统包括车载传感器模块、传感器驱动模块、地图模块、定位模块、感知模块、导航路径模块、融合决策模块、控制模块;本发明可实现对环岛复杂交互场景的适应,并保证深度强化学习算法训练过程中的安全决策,同时可实现强化学习算法的快速落地。
Intelligent vehicle roundabout decision-making rapid landing method and system based on reinforcement learning
基于强化学习的智能车辆环岛决策快速落地方法及系统
2024-01-23
Patent
Elektronische Ressource
Chinesisch
IPC: | G01C Messen von Entfernungen, Höhen, Neigungen oder Richtungen , MEASURING DISTANCES, LEVELS OR BEARINGS / G06N COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS , Rechnersysteme, basierend auf spezifischen Rechenmodellen / G08G Anlagen zur Steuerung, Regelung oder Überwachung des Verkehrs , TRAFFIC CONTROL SYSTEMS / H04W WIRELESS COMMUNICATION NETWORKS , Drahtlose Kommunikationsnetze |