首页|无人驾驶环境下的站台门与列车间安全全信息感知系统

无人驾驶环境下的站台门与列车间安全全信息感知系统

扫码查看
地铁全自动无人驾驶系统完全排除了司机和乘务人员,通过地面控制中心基于实时感知信息对列车和相关设施进行统一最优控制的自动运行系统.首先,对地铁全自动运行系统进行定义,并分析地铁站台门和轨行区侧的风险空间特点,提出在无人驾驶环境下对站台和轨行区侧进行安全信息感知的需求.其次,对现有信息感知技术进行评估,分析其优缺点.最后,提出一种基于顶装机器视觉的全时域全信息感知系统,能够实现列车进出站时客流量及异常行为的检测,站台门关闭后列车与站台门之间异物及异物种类的检测,以及列车不在站时站台门及轨行区异物的检测.该系统完全覆盖乘降作业监督区域,能够提供 1920×1080 分辨率或更高的实时图像,最小可检测的风险事件尺寸为 3 cm×3 cm.这些技术实现了站台门与列车间区域的全时空安全信息感知,避免了轨道交通系统功能碎片化,能够全面支撑未来智慧轨道交通信息发展的需求.
Research on Safety Total Information Awareness System Between Driverless Subway Platform Door and Train
The fully automated unmanned subway system eliminates the need for drivers and crew members,as it relies on real-time perception information from the ground control center to achieve unified and optimal control of trains and related facilities.This paper begins by providing a definition of the automated subway operation system and analyzes the specific risk characteristics of platform doors and track areas.It further identifies the need for safety information perception on the platform and track sides in the context of unmanned driving.Subsequently,an evaluation of existing information perception technologies is conducted,highlighting their strengths and weaknesses.Finally,a novel top-mounted machine vision-based comprehensive spatiotemporal information perception system is proposed.This system enables the detection of passenger flow and anomalous behaviors during train entry and exit,identifies foreign objects and their types between closed platform doors and trains,and detects foreign objects on the platform and track areas when the train is not present.The system provides full coverage of the supervision area for passenger boarding and alighting operations,delivering real-time images at a resolution of 1 920×1 080 or higher.It achieves a minimum detectable risk event size of 3 cm×3 cm.These cutting-edge technologies realize comprehensive spatiotemporal safety information perception between platform doors and trains,eliminating the fragmentation of functionality in rail transit systems and providing full support for the future development of intelligent rail transit information.

driverless subwaysafety information perceptionmachine visionpassenger flow detectionabnormal behavior detectionforeign object detection

王珩、廖先哲、刘伟铭、戴愿、杨代鑫、刘一霄

展开 >

深圳市地铁集团有限公司,深圳 518026

华南理工大学土木与交通学院,广州 510641

广州顺如电子科技有限公司,广州 510645

无人驾驶地铁 安全信息感知 机器视觉 客流量检测 异常行为检测 异物检测

2025

铁道标准设计
中铁工程设计咨询集团有限公司

铁道标准设计

北大核心
影响因子:1.129
ISSN:1004-2954
年,卷(期):2025.69(1)