Principal Investigator/Researcher
Yingtao Jiang, Hualing (Harry) Teng, Tianding Chen and Han Li, University of Nevada, Las Vegas
Project Description
In this study, we propose to develop and test a platform that utilizes an image sensor (camera) and a LIDAR sensor mounted a multi-rotor UAV for autonomous and efficient monitoring and measurement of rail track’s longitudinal and other geometrical irregularities. Flying at a low altitude, normally around 10 meters above the ground level, the UAV during one single flight that lasts from 15 to 25 minutes is expected to photograph and optically scan 5+ miles of rail tracks. The images and the light reflection time signals thus collected, together with the track’s geospatial position determined by the GPS and the UAV’s state (velocity, acceleration, yaw, etc. by IMU, are fused altogether to build the point cloud model of the tracks that have been inspected by the UAV. With sub-cm accuracy level, this point cloud model obtained is compared off line against the baseline track data established from the previous measurements and/or the design specifications stored in the track rail database. If the results indicate existence of possible problems like track subsidence, deformation and component damage, a second inspection and immediate maintenance service will then be warranted. Note that the UAV is given the capability, with the help of the LIDAR and proximity sensors, to navigate through tunnels or any other uncertain waypoint paths. When passing a tunnel or flying in the dark, the UAV may have to fly at a lower altitude and a lower speed, and the on-board LED lighting device may have to be automatically turned on and the light beams will be focused onto the tracks for imaging purposes.
Final Report
Outputs:
- Han Li, Tianding Chen, Hualiang Teng, Yingtao Jiang, A graph-based reinforcement learning method with converged state exploration and exploitation, Computer Modeling in Engineering & Sciences, Volume 118, Issue 2, Pages 253-274, 2019
- Lihao Qiu, Development of UAV-Based Rail Track Geometry Irregularity Monitoring and Measuring Platform Empowered by Artificial Intelligence, Thesis of Electrical and Computer Engineering at the University of Nevada Las Vegas, 2022