Range Only Simultaneous Localization and Mapping Using Paired Comparisons

Range-only Simultaneous Localization and Mapping (SLAM) considers the problem of tracking a moving target along with estimating locations of fixed landmark points in an environment. Most prior works solve the problem based on the direct knowledge of a sequence of distance measurements between the target and the landmarks. We consider the scenario where the exact distance measurements are either unavailable or unreliable due to high noise levels. Instead, we assume the knowledge of a sequence of relative distance measurements and initial location estimates for the landmarks. Specifically, we assume that the measurements are available in the form - t is closer to a than b - where t refers to the target and a, b is a pair of landmarks. We propose a particle filter-based Bayesian method to simultaneously localize the target and map the landmarks using such paired comparisons. The preliminary results of the proposed method are presented by performing simulations. A potential application of this work includes RFID-based object localization where a prior exact knowledge of the map is not available.