Algorithms for Robot Navigation
CS
480: Robotics & 3D Printing Lecture, Dr. Lawlor
Iterative
Closest Point: match up two point clouds by aligning each
point with the closest point on the other mesh. Useful for
correlating point clouds from Kinect or other depth scanners.
(Free implementation in MeshLab.)
Random Sample
Consensus (RANSAC): avoid outliers, by randomly picking points
until you find a set that is self-consistent.
Simultaneous
Localization and Mapping (SLAM): match up series of LIDAR
scans using RANSAC.