The invention relates to an emergency braking anti-collision control method for a subway train. The method comprises the following steps: obtaining road condition pictures of no foreign matters and foreign matters in front of a train during actual normal driving, and constructing a road condition detection model based on a YOLOv4 target recognition network; deploying the road condition detection model in a train emergency braking collision avoidance system to monitor and output current road condition data in real time; and according to the current road condition data, in combination with a distance signal output by a train radar range finder, obtaining a real-time braking safety distance, outputting a corresponding early warning signal, and outputting a corresponding control signal to a brake. Compared with the prior art, the method is independent of an automatic train control system, is high in specificity and stability, is based on the convolutional neural network and is fused with a radar ranging technology, so that the current road condition is accurately detected, the real-time braking safety distance is determined, and anti-collision early warning and emergency braking can be reliably carried out.
本发明涉及一种地铁列车紧急制动防撞控制方法,包括:获取列车实际正常行驶时前方无异物和有异物的路况照片,并基于YOLO v4目标识别网络,构建得到路况检测模型;将路况检测模型部署在列车紧急制动防撞系统中,以实时监测输出当前路况数据;根据当前路况数据,结合列车雷达测距仪输出的距离信号,得到实时制动安全距离,并输出相应预警信号、输出相应控制信号给制动器。与现有技术相比,本发明独立于列车自动控制系统,专一性强、稳定性高,本发明基于卷积神经网络并融合雷达测距技术,以准确检测当前路况、并确定实时制动安全距离,能够可靠地进行防撞预警及紧急制动。
Emergency braking anti-collision control method for subway train
一种地铁列车紧急制动防撞控制方法
15.02.2022
Patent
Elektronische Ressource
Chinesisch
Auxiliary anti-collision system and method suitable for subway train
Europäisches Patentamt | 2022
|Train external emergency braking system and emergency braking method
Europäisches Patentamt | 2020
|