Robust Real-Time Local Laser Scanner Registration with Uncertainty Estimation.

We presenta fast, robust method for registering successive laser rangefinder scans. Correspondences between the current scan and previous scans are determined. Gaussian uncertainties of the correspondences are generated from the data, and are used to fuse the data together into a unified egomotion estimate using a Kalman process. Robustness is increased by using a RANSAC variant to avoid invalid point correspondences. The algorithm is very fast; computational and memory requirements are O(nlogn) where n is the number of points in a scan. Additionally, a covariance suitable for use in SLAM and filter techniques is cogenerated with the egomotion estimate. Results in large indoor environments are presented.