Improved motion tracking with velocity update and distortion correction from planar laser scan data

摘要: A rigid-body transformation can be estimated from the changes of observed surrounding by comparing two laser scan data that are captured at different time step. An estimated transformation might be used for aligning two different scan data to be seamlessly connected and/or tracking the motion of the sensor itself. Laser scan data consists of a number of distance value, and the distance values in the same scan data are not scanned at the same time which means that each distance value has significant time difference in its nature. Laser scan data might be distorted when sensor moves during scanning because of its time difference. In this paper, we suggest a methodology to compensate the distortion by estimating the velocity of sensor and adopting the velocity to scan data to be aligned at the same time. In addition, outliers are effectively rejected during the iteration of velocity update. Finally, more accurate and …