The invention provides a high-speed train formation anti-collision control method based on deep learning. The method comprises the following steps: sensing multi-dimensional running state information of a train in real time based on a multi-sensor fusion technology, and realizing real-time interaction of state information of adjacent trains based on a train-train communication technology; based on the relative braking mode, establishing a virtual reconnection formation safety interval model, and further calculating an ideal tracking interval; for the virtual reconnection pilot train, an operation optimization curve is calculated off line based on a genetic algorithm; for a virtual reconnection following train, the problem of inaccurate prediction positioning of a pilot train is solved based on a full-connection neural network, and a prediction result is applied to a model prediction controller of the following train to realize formation cooperative control. From two aspects of overspeed protection and collision prevention, collision risks are evaluated in real time, and braking measures are adjusted in time. According to the invention, more accurate speed following and safe and stable short-distance tracking of the high-speed train in the virtual reconnection tracking mode can be realized.

    本发明提供了一种基于深度学习的高速列车编队防碰撞控制方法。该方法包括:基于多传感器融合技术实时感知列车多维运行状态信息,基于车‑车通信技术实现相邻列车状态信息的实时交互;基于相对制动模式,建立虚拟重联编队安全间隔模型,进而计算理想追踪间隔;针对虚拟重联领航列车,基于遗传算法离线计算运行优化曲线;针对虚拟重联跟驰列车,基于全连接神经网络解决对领航列车的预测定位不准确问题,将预测结果应用于跟驰列车的模型预测控制器中实现编队协同控制;从超速防护和碰撞预防两方面着手,实时评估碰撞风险并及时调整制动措施。本发明能实现高速列车虚拟重联追踪模式下更精准的速度跟随及安全稳定的短距追踪。


    Access

    Download


    Export, share and cite



    Title :

    High-speed train formation anti-collision control method based on deep learning


    Additional title:

    一种基于深度学习的高速列车编队防碰撞控制方法


    Contributors:
    SHANG GUANWEI (author) / ZHU YUENING (author) / DUN YICHEN (author) / CHAI LINGUO (author) / ZHU ZILE (author) / WANG WEIJIE (author)

    Publication date :

    2024-02-27


    Type of media :

    Patent


    Type of material :

    Electronic Resource


    Language :

    Chinese


    Classification :

    IPC:    B61L Leiten des Eisenbahnverkehrs , GUIDING RAILWAY TRAFFIC




    COOPERATIVE FORMATION TRAIN ANTI-COLLISION METHOD AND DEVICE

    ZHANG QIANG / SUN JUNGUO / WEN YILONG | European Patent Office | 2023

    Free access

    COOPERATIVE FORMATION TRAIN ANTI-COLLISION METHOD AND DEVICE

    ZHANG QIANG / SUN JUNGUO / WEN YILONG | European Patent Office | 2021

    Free access

    DEEP LEARNING - BASED STOP CONTROL METHOD AND SYSTEM FOR HIGH-SPEED TRAIN

    YIN JIATENG / NING CHENHE / SU SHUAI et al. | European Patent Office | 2022

    Free access

    Whole train dynamic and stability in high speed collision

    Lupton, Jim / Bound, Sturat M. | IuD Bahn | 2006