Laser pose tracking for a mobile robot using fuzzy adaptive extended information filtering

Abstract

This paper presents a laser pose tracking method for a mobile robot using fuzzy adaptive extended information filtering (FAEIF). The FAEIF, composed of a fuzzy tuner, is proposed to improve the performance of the EIF, and to avoid the EIF divergence problems caused by the noise variations or occurrence of modeling errors. By monitoring the innovation sequences, the fuzzy tuner adjusts on-line the key parameter of the EIF and thus improves accuracy and robustness of its estimates. Through simulation and experiment results, the proposed laser-scanner-based pose estimation method has been proven capable of giving satisfactory performance and overcoming the filter divergence problems under the noise variations and occurrence of modeling errors.