Robust SVSF-SLAM for Unmanned Vehicle in Unknown Environment

Abstract

Simultaneous localization and mapping (SLAM) is an important topic in the autonomous mobile robot research. The most popular solutions of this problem are the EKF-SLAM and the FAST-SLAM, the former requires an accurate process and observation model and suffer from the linearization problem, and the latter is not suitable for real time implementation. Therefore, a new alternative solution based on the smooth variable structure filter (SVSF-SLAM) algorithm is proposed in this paper to solve the Unmanned Ground Vehicle (UGV) SLAM problem. The SVSF filter which is formulated in a predictor-corrector format is robust face parameters uncertainties and error modeling and doesn’t require any assumption about noise characteristics. In this paper the SVSF-SLAM algorithm is implemented using the odometer and LASER data to construct a map of the environment and localize the UGV within this map. The proposed algorithm is validated and compared to the EKF-SLAM algorithm. Good results are obtained by the SVSF-SLAM comparing to the EKF-SLAM especially when the noise is colored or affected by a variable bias. Which confirm the efficiency of our approaches.