Monitoring rail infrastructure using multisensor navigation on a moving platform and autonomous robots

摘要

RailSat aims to use Global Navigation Satellite System (GNSS) to monitor and maintain railway assets and its surrounding environment by railway asset owners and/or other relevant stakeholders.udThe rail sector is looking for continuous monitoring solutions which have no impact on the train service, both wayside (track bound) and onboard (train bound), which require accurate positioning while travelling at high speeds (>120kmh).udThis paper focuses on the combination of positioning data from traditional GNSS/INS system with processed LIDAR point cloud and discusses real-life results from the Snake Pass, Peak District, England. Data have been collected using a dedicated multisensory van but the nature of the road allows us to draw conclusions relevant to the rail industry.udThis paper discusses the proposed deployment of a mobile LiDAR monitoring system consisting of a set of laser scanners and a navigation component. While the LIDAR component is capable of centimetre accuracy, it is limited by the navigation accuracy, predominantly affected by the difficult railway environment, frequent multipath and NLOS interference combined with a loss of signal next to the monitoring structures itself (bridges, cuttings, tunnels, embankments etc.), making precise positioning the biggest challenge.udThe proposed navigation system combines IMU positioning system with a computer vision system capable of localisation using features in the natural environment. This paper outlines the combination of the proposed navigation system with the LIDAR’s information, which provides two ways of correcting navigation trajectory in post-processing.