Maglev Obstacle Detection System Based on Laser Radar

Abstract

Magnetic levitation train is a new type of rail transit system which uses electromagnetic force to realize traction drive. Monitoring the maglev track and judge whether obstacles intrude into the vehicle safety limitation is very important to train safe operation. Vehicle-borne laser scanning technology has been gradually applied to the deformation detection of highway, tunnel and other scenes. In this paper, we designed a maglev mobile scanning detection system based on multi-sensors. Three laser profile scanners were used to collect track point cloud data, a GPS clock of the NTP protocol and the GPS time, Pulse Per Second (PPS) signal from an inertial measurement unit (IMU) were respectively provided for time synchronization of SICK and Z+F, the IMU and GNSS antenna were combined as positioning and attitude (POS) system. Using the calculated spatial conversion information, the acquisition data of multi-sensors were geo-referenced by coordinate transformation. Using the obtained three-dimensional (3D) continuous point cloud of the maglev track, the original reference data of obstacle-free is indexed by K-D tree as the initial map. For the point cloud to be detected, query point and query distance threshold are set to carry out the range search and feature matching, the point cloud clustering of the obstacle are collected. Finally, the independent obstacles are output by the processing of European clustering. Through the test of maglev data acquisition experiment and obstacle simulation experiment, the accuracy of 3D point cloud is verified, and the detection of obstacles are finished.