The invention provides an automatic driving vehicle interaction control method based on reinforcement learning, and the method comprises the steps: obtaining data for model training, and also comprises the following steps: carrying out the pre-training of a decision network; formulating a reward function of reinforcement learning; training a reinforcement learning model; and learning model testing is carried out. The invention provides an automatic driving vehicle interaction control method based on reinforcement learning, supervised learning and deep reinforcement learning are combined, human driving experience is added into a learning sample, the human driving experience is utilized in the learning process to enable an unmanned vehicle to learn an optimal strategy, and meanwhile, an adaptive balance sampling method is adopted, so that the self-adaptive balance sampling method is utilized to realize the self-adaptive self-adaptive self-adaptive self-adaptive self-adaptive self-adaptive self-adaptive self-adaptive self-adaptive self-adaptive self-adaptive self-adaptive self-adaptive self The automatic driving vehicle automatically adjusts the sampling ratio of self-exploration experience and human driving experience in different learning stages, and a reward function is designed for different evaluation indexes, so that the automatic driving vehicle is further guided to better learn an interaction control strategy.
本发明提供基于强化学习的自动驾驶车辆交互控制方法,包括获取用于模型训练的数据,还包括以下步骤:进行决策网络的预训练;制定强化学习的奖励函数;强化学习模型训练;进行学习模型测试。本发明提出了基于强化学习的自动驾驶车辆交互控制方法,将监督学习和深度强化学习相结合并在学习样本中加入了人类驾驶经验,学习过程中利用了人类驾驶经验使无人驾驶车辆学习最优策略,同时采用自适应平衡采样方法,使自动驾驶车辆在不同学习阶段自动调整自我探索经验和人类驾驶经验的采样占比,并针对不同评判指标设计了奖励函数,进一步引导自动驾驶车辆更好地学习交互控制策略。
Automatic driving vehicle interaction control method based on reinforcement learning
基于强化学习的自动驾驶车辆交互控制方法
2024-09-17
Patent
Electronic Resource
Chinese
IPC: | G06N COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS , Rechnersysteme, basierend auf spezifischen Rechenmodellen / 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 |
Automatic driving vehicle control method based on multi-agent reinforcement learning
European Patent Office | 2023
|Automatic driving vehicle microscopic decision-making method based on reinforcement learning
European Patent Office | 2020
|Automatic driving vehicle decision-making method based on implicit time scale reinforcement learning
European Patent Office | 2025
|Automatic driving vehicle lane changing method based on double-layer deep reinforcement learning
European Patent Office | 2023
|Automatic driving vehicle high-speed ramp intelligent afflux method based on reinforcement learning
European Patent Office | 2023
|