Sum of Gaussian Filter based Cooperative Range-Only SLAM

Abstract

In most range-only simultaneous localization and mapping (RO-SLAM) algorithms, only direct measurements between robot and its neighbor nodes have been utilized for the localization, even though the localization accuracy can be improved if other inter-beacon measurements are actively employed. In this paper, a sum of Gaussian (SoG) filter based RO-SLAM is proposed, whereby all inter-node measurements are integrated for the localization. The performance of the proposed algorithm is verified with the real experimental data, which results in a smaller map estimation error with a faster convergence speed than the conventional Rao-Blackwellized particle filter (RBPF) based RO-SLAM without using the inter-beacon measurements.