The invention relates to an automatic driving vehicle high-speed ramp intelligent convergence method based on reinforcement learning, and belongs to the field of automatic driving automobiles, and the method comprises the following steps: S1, constructing a reinforcement learning intelligent convergence model based on a strategy gradient of depth certainty, including constructing a state space, determining an action space, and constructing a reward function; giving structures of a strategy function and an evaluation model; s2, using a two-degree-of-freedom automobile kinematics model to describe the motion of the vehicle in the training scene, and using a rule-based intelligent decision planning model to control the vehicle in the environment, so that the environment has interaction characteristics; and S3, training the reinforcement learning intelligent convergence model of the strategy gradient based on depth certainty. According to the method, the optimality, the stability and the high efficiency of the decision planning strategy of the autonomous vehicle in the high-speed ramp convergence scene are improved by using the reinforcement learning algorithm.
本发明涉及一种基于强化学习的自动驾驶车辆高速匝道智能汇入方法,属于自动驾驶汽车领域,包括以下步骤:S1:构建基于深度确定性的策略梯度的强化学习智能汇入模型,包括搭建状态空间,确定动作空间,构建奖励函数,给定策略函数和评价模型的结构;S2:使用二自由度汽车运动学模型描述训练场景中车辆的运动,并利用基于规则的智能决策规划模型控制环境车辆,使环境具有交互特性;S3:训练所述基于深度确定性的策略梯度的强化学习智能汇入模型。本发明利用强化学习算法提高了自动驾驶汽车在高速匝道汇入场景下决策规划策略的最优性、稳定性与高效性。
Automatic driving vehicle high-speed ramp intelligent afflux method based on reinforcement learning
一种基于强化学习的自动驾驶车辆高速匝道智能汇入方法
2023-12-15
Patent
Electronic Resource
Chinese
IPC: | B60W CONJOINT CONTROL OF VEHICLE SUB-UNITS OF DIFFERENT TYPE OR DIFFERENT FUNCTION , Gemeinsame Steuerung oder Regelung von Fahrzeug-Unteraggregaten verschiedenen Typs oder verschiedener Funktion |
Vehicle ramp afflux-in and afflux-out cooperative control method
European Patent Office | 2022
|Vehicle ramp afflux method, device and equipment and storage medium
European Patent Office | 2023
|DQN and CNN-based expressway automatic driving exclusive lane vehicle afflux method
European Patent Office | 2024
|Vehicle afflux decision-making method considering equality of afflux motivation
European Patent Office | 2023
|