Map alignment based on PLICP algorithm for multi-robot SLAM

Abstract

Inter-robot observation (Rendezvous) strategy is usually adopted to align local maps in multi-robot SLAM system while the initial relative positions are unknown. However, the accuracy of the estimated alignment is often affected by various uncertainty sources. This paper presents a new approach based on the Point-to-Line Iterative Closest Point (PLICP) algorithm to improve the accuracy of the map alignment: First, local maps are generated by means of the FastSLAM algorithm, then the initial alignment parameters are calculated by inter-robot observation, and finally the PLICP algorithm is used to update these parameters. Experimental results illustrate the accuracy improvement of the proposed approach.