Directional Endpoint-based Enhanced EKF-SLAM for Indoor Mobile Robots

This paper proposes an enhanced Extended Kalman Filter (EKF)-based Simultaneous Localization and Mapping (SLAM) algorithm based on ‘directional endpoint’ features extracted from two-dimensional (2D) laser data for indoor environments. The proposed approach is composed of calculating the covariances of the extracted line segments, calculating the covariances of the directional endpoints, and enhanced EKF-SLAM. Different from the classical SLAM based on point and line features, this work uses the directional endpoint feature, which has 3 degrees of freedom. To facilitate the enhanced EKF-SLAM, the implicit function theorem and the geometrical method are used to obtain the uncertainty of the directional endpoint. Comparative experimental results show superior performance of our proposed algorithm. In addition, the enhanced EKF-SLAM achieves the similar performance compared with Karto-SLAM in terms of pose estimation, but at the same time, the feature map composed of a set of directional endpoints is obtained, which is robust in dynamic environments.