Simultaneous Localization and Mapping Using Invariant Natural Features

Abstract

This paper presents a practical approach to solve mobile robot simultaneous localization and mapping (SLAM) problem using natural visual landmarks obtained from a monocular vision. The Rao-Blackwellised particle filter (RBPF) is used to extend the path posterior by sampling new poses that integrate the current observation. The landmark position estimation and update is implemented through extended Kalman filter (EKF). Furthermore, the number of resampling steps is determined adaptively, which seriously reduces the particle depletion problem. Single CCD camera tracks the 3D natural landmarks, which are structured with matching image features extracted through Scale Invariant Feature Transform (SIFT). The matching for highly distinctive SIFT features described with multi-dimension vector descriptor is implemented with a KD-Tree in the time cost of O(log2N). And the matches with large error are eliminated by epipolar line constraint approach. The dense metric maps of natural 3D point landmarks for indoor environments is constructed. Experiments on the robot Pioneer3 in our real indoor environment indicate superior performance.