Map building of unknown environment using PSO-tuned enhanced Iterative Closest Point algorithm

Abstract

Iterative Closest Point (ICP) algorithm is widely used in 2D and 3D spatial and geometric alignment. There are many variants of the ICP algorithm, proposing methods to minimize the sum of Euclidean distances between two clouds of scanning points for map building of an unknown environment by a mobile robot. Considering simplicity and computational efficiency, this paper proposes an enhanced-ICP incorporating a Particle Swarm Optimization (PSO) to effectively filter out outliers and avoid the false matching points during the map building process. Experimental results showed that, the proposed PSO-tuned enhanced-ICP can effectively reduce the accumulated errors to improve the map building accuracy by circumventing the problems of local optimal solutions resulted from the outliers and false matching points during the map building process.