As one of the most important public transportation, urban rail transit bears great operational pressure and relies completely on signaling system to maintain its safety and efficiency. In order to improve train operation efficiency and safety, train active obstacle perception is a good solution. In this paper, we introduce the train intelligent detection system (TIDS) and its application in practical line. TIDS consists of data pre-processing layer, perception layer, and decision-making layer. The data pre-processing layer mainly includes data acquisition and spatio-temporal synchronization. The perception layer applied convolutional neural network to achieve rail area detection and train detection simultaneously in each camera image, and further combine lidar points to measure the train distance. And the decision-making layer judge the relationship between detected trains and the rail area, so as to determine whether the detected trains affect the normal running of the train. At this stage, TIDS system has been comprehensively tested on Beijing Yanfang Line, Shanghai Line 6, Chengdu Line 3, Hong Kong Tsuen Wan, etc. It has been tested and accumulated more than 2 TB data. In addition, we analyze the running accuracy of TIDS system in detail, the results showed that the leak detected rate (LDR) is 0.007 and the mistakenly detected rate (MDR) is 0.000478. Particularly, the system is being practical applied in Hong Kong Tsuen Wan line.


    Access

    Check access

    Check availability in my library

    Order at Subito €


    Export, share and cite



    Title :

    Application of Train Intelligent Detection System in Large-Volume Operated Lines


    Contributors:

    Conference:

    13th Asia Pacific Transportation Development Conference ; 2020 ; Shanghai, China (Conference Cancelled)



    Publication date :

    2020-06-29




    Type of media :

    Conference paper


    Type of material :

    Electronic Resource


    Language :

    English



    Application of Train Intelligent Detection System in Large-Volume Operated Lines

    Wang, Zhangyu / Yang, Zining / Tang, Simon et al. | TIBKAT | 2021


    COMBAT - A Train Control System for Secondary Lines by Radio-Based Train Detection

    Matsuzawa, T. / Nishibori, N. / Sasaki, T. et al. | British Library Online Contents | 2004




    Intelligent train detection robot system and detection method thereof

    LI TAO / WANG MINQI / YANG JUN et al. | European Patent Office | 2022

    Free access