Incremental closed-form solution to globally consistent 2D range scan mapping with two-step pose estimation

Abstract

This paper presents a simplification of the Lu and Milios work [1] for globally consistent Simultaneous Localization And Mapping (SLAM) with two-dimensional (2D) scan registration. The proposed method is based on global two-step (orientation-position) pose estimation. This approach reduces the computational effort since it represents a closed-form solution that admits an incremental updating procedure. Successful simulation and experimental tests with the tracked mobile robot Auriga-¿ with an onboard laser rangefinder in indoors are presented.