The invention discloses an automatic driving solution under a multi-target complex traffic scene based on reinforcement learning, which can process all traffic scenes by using a set of reinforcement learning automatic driving modeling method and has better universality. The reinforcement learning comprehensive modeling is based on a traditional reinforcement learning framework, and environment perception information and feature quantity extracted in combination with human knowledge are used as observation space. Model training is based on a time-varying training strategy, and the training speed and the generalization of strategy application are improved. In order to further guarantee the form safety of the vehicle, a dangerous action recognizer based on a long-short term memory (LSTM) network and a rule constraint device based on a human knowledge body are provided, the dangerous action recognizer is sampled from the environment and trained, so that the vehicle has the ability of recognizing dangerous actions and dangerous scenes, and the safety of the vehicle is improved. And rule constraints are designed for specific situations to limit output actions, so that the safety can be greatly improved, the collision frequency is reduced, and the driving safety of the vehicle is guaranteed.

    本发明公开一种基于强化学习的多目标复杂交通场景下自动驾驶解决方法,该方法可以使用一套强化学习自动驾驶建模方法处理所有交通场景,具有较好的通用性。强化学习综合建模基于传统强化学习框架,使用环境感知信息及结合人类知识提取的特征量作为观测空间。模型训练基于时变训练策略,提高训练速度和策略应用的泛化性。为对其形式安全性作进一步保障,还提出了基于长短时记忆(LSTM)网络的危险动作识别器与基于人类知识体的规则约束器,从环境中采样并训练危险动作识别器,使车辆具备识别危险动作与危险场景的能力,并针对特定情形设计规则约束对输出动作加以限制,可以大大提高安全性,减少碰撞次数,以保障车辆的行驶安全。


    Access

    Download


    Export, share and cite



    Title :

    Automatic driving solution under multi-target complex traffic scene based on reinforcement learning


    Additional title:

    基于强化学习的多目标复杂交通场景下自动驾驶解决方法


    Contributors:
    CHI YUXIANG (author) / FAN YU (author)

    Publication date :

    2022-07-05


    Type of media :

    Patent


    Type of material :

    Electronic Resource


    Language :

    Chinese


    Classification :

    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 / G06N COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS , Rechnersysteme, basierend auf spezifischen Rechenmodellen



    Automatic driving decision-making method and system in complex traffic scene based on reinforcement learning

    WU ZHIFEI / ZHANG SHAOJIE / WU XIN et al. | European Patent Office | 2023

    Free access

    Intelligent automobile reinforcement learning cooperative driving method for urban traffic scene

    LIU JIA / LIANG QINGYI / YIN JIANWEN et al. | European Patent Office | 2024

    Free access

    Automatic driving method based on digital traffic scene

    FAN TING / NIE JIN | European Patent Office | 2021

    Free access

    Automatic driving vehicle control method based on multi-agent reinforcement learning

    MA HONGBIN / LIU PING / JIN YING et al. | European Patent Office | 2023

    Free access