At present, an AEB system cannot avoid collision under the emergency situation that the distance between the AEB system and an obstacle is smaller than the total longitudinal braking distance, and only the same braking action can be performed when the AEB system faces different obstacles only by means of sensor information, so that the invention provides an automobile emergency collision avoidance control method based on DQN deep reinforcement learning. Belongs to the field of new energy automobile braking, images and sensor information are spliced to serve as state input, a transverse avoidance action is added on the basis of longitudinal braking, the problem that the collision avoidance effect is poor due to the fact that an AEB system only depends on longitudinal braking in an emergency is solved, the braking behavior is more targeted when facing different obstacles, and the collision avoidance efficiency is improved. The method comprises the steps of subtask design, state and action space design, multi-target reward function design and DQN parameter setting and training. According to the method, the algorithm training efficiency is improved, the safety of the automobile is improved, and the avoidance strategy of the automobile is more humanized.

    目前AEB系统在与障碍物距离小于总纵向制动距离的突发情况下无法避免碰撞,且只依靠传感器信息在面对不同的障碍物时,只能做出相同的制动动作,因此,本发明提出一种基于DQN深度强化学习的汽车紧急避撞控制方法,属于新能源汽车制动领域,将图像与传感器信息拼接作为状态输入,在纵向制动基础上加入横向避让动作,拟解决AEB系统在突发情况下仅依靠纵向制动避撞效果不好的问题,以及在面对不同障碍物时的制动行为更具有针对性,该方法包括子任务设计、状态与动作空间设计、多目标奖励函数设计、DQN参数设置及训练;本发明提高了算法训练效率,提高了汽车的安全性,使汽车的避让策略更加人性化。


    Access

    Download


    Export, share and cite



    Title :

    Automobile emergency collision avoidance control method based on DQN deep reinforcement learning


    Additional title:

    一种基于DQN深度强化学习的汽车紧急避撞控制方法


    Contributors:
    LU XIAOHUI (author) / ZHENG XINYI (author) / LYU XINZHAN (author) / LI SHAOSONG (author) / LI JIACHUN (author) / DONG XUSHENG (author) / ZHANG PENGFEI (author) / ZHANG NA (author)

    Publication date :

    2023-05-30


    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



    Vehicles Control: Collision Avoidance using Federated Deep Reinforcement Learning

    Elallid, Badr Ben / Abouaomar, Amine / Benamar, Nabil et al. | ArXiv | 2023

    Free access

    Pedestrian Collision Avoidance Using Deep Reinforcement Learning

    Rafiei, Alireza / Fasakhodi, Amirhossein Oliaei / Hajati, Farshid | Springer Verlag | 2022