A mobile robot exploring an unknown environment may need to keep track of its pose through its sensors. Global scan matching is a way of computing the pose difference of a robot at two different locations on the navigation path by finding common features observed in range sensor readings recorded at these locations. In this thesis, we introduce a line segment based global scan matching algorithm which computes the pose difference by matching common line segments extracted from two laser range scans from different unknown poses. Matching is performed by exploiting the invariant geometry, which we call geometrical relations, among line segments. The use of geometrical relations instead of points also reduces the computational complexity of determining the pose difference between two distinct scans. Compared to other scan matching algorithms, our method presents a powerful means for map building, place recognition, loop closing and multirobot mapping, all in real-time.

DATE: 29 July, 2008, Tuesday@ 13:30
PLACE: EA 502