ArUco-based Automatic Extrinsic Sensor Calibration for Driverless Train System*

To realize driverless operation of railways on general lines, highly reliable forward monitoring devices are required. To ensure high reliability, a sensor fusion that integrates the detection results of multiple sensors is used, but in order to integrate properly, strict calibration of each sensor position and posture is required. In this research, a target based automatic extrinsic sensor calibration algorithm to obtain the relative pose transformation between camera and 3D Light Detection And Ranging (LiDAR) is proposed. The fiducial marker called ‘ArUco’ markers are used as a calibration target. The method employs ArUco marker feature points (corners as a feature) extraction simultaneously by the camera and LiDAR sensor data. Determining the relationship between the coordinates of the features of two data sets, a least square method with Single Value Decomposition (SVD) method for matrix formulation is utilized.