Effective Line-Based SLAM with Adaptive Tuning of Particles

Abstract

This paper deals with the fast SLAM algorithm for the line-based SLAM problem with a laser range scanner for a single two-wheeled mobile robot. Since the computational time of the estimation process per each step depends on the number of observed landmarks, and we control the computational time by adaptively tuning the number of particles according to the number of the observed landmarks. First, we review the estimation process of the fast line-based SLAM algorithm. Then, we propose a method for the prediction of the computational time and how to control it by using the number of particles. Finally, we show simulation results of the proposed method in order to verify its effectiveness.