The invention provides an energy feedback type active suspension control method based on a multi-agent reinforcement learning algorithm, and the method comprises the steps: building a kinetic model of an energy feedback type active suspension, and building a random pavement excitation model as the input of an energy feedback type active suspension system, then, a multi-agent reinforcement learning reward function is designed based on the energy feedback type active suspension dynamical model, a multi-agent reinforcement learning controller is built based on the energy feedback type active suspension dynamical model and the multi-agent reinforcement learning reward function, and through continuous interaction between the multi-agent reinforcement learning controller and the energy feedback type active suspension, the energy feedback type active suspension is obtained; and updating network parameters of the multi-agent reinforcement learning control controller to obtain an optimal control strategy, and finally controlling the two actuators based on the optimal control strategy obtained by updating the parameters of the controller so as to control the energy feedback type active suspension. According to the control method for the energy feedback type active suspension, the comfort performance and the energy consumption characteristic of the vehicle are considered at the same time, the robustness and accuracy of energy feedback type active suspension control can be improved, the vehicle riding comfort can be improved, and the energy consumption of the suspension in the driving process can be reduced.

    本发明提供了一种基于多智能体强化学习算法的馈能式主动悬架的控制方法,该方法包括:搭建馈能式主动悬架的动力学模型以及搭建随机路面激励模型作为馈能式主动悬架系统的输入,然后基于馈能式主动悬架动力学模型设计多智能体强化学习奖励函数并且基于馈能式主动悬架动力学模型以及多智能体强化学习奖励函数搭建多智能体强化学习控制器,通过多智能体强化学习控制器与馈能式主动悬架的不断交互,更新多智能体强化学习控制控制器的网络参数得到最优控制策略,最后基于更新控制器参数得到的最优控制策略控制两个作动器进而控制馈能式主动悬架。通过本申请,实现了对于车辆同时兼顾舒适性能和能耗特性的馈能式主动悬架的控制方法,能够提升馈能式主动悬架控制的鲁棒性、精确性,还能够提升车辆乘坐舒适度以及减小行驶过程当中悬架的能耗。


    Access

    Download


    Export, share and cite



    Title :

    Energy feedback type active suspension control method based on multi-agent reinforcement learning control


    Additional title:

    基于多智能体强化学习控制的馈能式主动悬架控制方法


    Contributors:
    GAO XIAOTIAN (author) / DU YU (author) / HAN SHIYUAN (author) / ZHONG YIXIN (author) / CAO SHUYU (author) / LIAN YANG (author)

    Publication date :

    2025-04-15


    Type of media :

    Patent


    Type of material :

    Electronic Resource


    Language :

    Chinese


    Classification :

    IPC:    B60G VEHICLE SUSPENSION ARRANGEMENTS , Radaufhängungen und Federungen für Fahrzeuge



    Linear motor energy feedback type active suspension actuator and control method

    KAGEYAMA KEI / WU ZHENGFEI / SUN YAHONG et al. | European Patent Office | 2024

    Free access

    Active Suspension Control Based on Multi-agent Predictive Algorithm*

    Zhang, Niaona / He, Yang / Whang, Yang et al. | IEEE | 2020


    Energy feedback type magneto-rheological semi-active suspension and control method thereof

    ZHOU CHENYU / ZHAO XUAN / YU QIANG et al. | European Patent Office | 2024

    Free access

    Composite energy feedback type vehicle semi-active suspension actuator and control method thereof

    KOU FARONG / HAO SHUAISHUAI | European Patent Office | 2020

    Free access

    Semi-active hydraulic suspension multi-objective optimization reinforcement learning control method based on T-S fuzzy

    HAN SHIYUAN / HE GUANGMING / LIANG TONG et al. | European Patent Office | 2024

    Free access