Spherical terrain matching for SLAM in planet exploration

This paper describes a scan matching algorithm for motion estimation near a planet surface using scanning laser scanner data in spherical coordinate. It is directly based on range finding data and followed by point to point terrain map alignment in the laser scanner's spherical coordinate system. Laser scan matching of current and reference scans are enhanced by weighted terrain and distortion compensation. It is also accelerated by predicted vision window using inner dynamic model and SLAM results. The algorithm is tested using data acquired within virtual OpenGL environment and proved to be efficient for scan matching with terrain distortion.