A novel railway maintenance robot for inspection and repair

Robotics and automation are widely used in various sectors for their economic benefits, accuracy, and efficiency. However, the railway industry has been slow to adopt these technologies for track inspection and repair, despite the increasing demands of asset management. Advances in robotics, computing, sensors, and Industry 5.0 are driving companies to explore innovative inspection and repair methods to optimize resource usage. Mobile manipulators, which combine robot mobility with industrial precision, have the potential to replace humans in risky and tedious tasks. In this research, a Robotic Inspection and Repair System (RIRS) has been developed to establish an improved track inspection method and a robotic repair technique. With a 0.27% error rate in calculating defect positions and 1mm precision in actuation, RIRS demonstrates strong feasibility for railway track maintenance. Later, an improved inspection method has been proposed fusing 3D reconstructed model of the target object from the monocular camera with the large-scale Global Positioning System (GPS) data, medium-scale environment perception from Light Detection And Ranging (LiDAR) and small-scale Colour and Depth (RGB-D) model. Texture of surface and less than 5%-dimensional error of the 3D model compared to the physical model ensures the credibility of the improved inspection technique which provides more information of the target object and sets the foundation for a data-rich digital twin in the future. Finally, the framework of both human-in-the-loop repair task and autonomous simulated repair task have been proposed. Successful delivery of the correct tool based on the detected defect upon receiving command from the human operator and automatic circular trajectory generation demonstrate the prospects of RIRS for assisting human in track repair tasks.