Pose-Sequence-Based Graph Optimization Using Indoor Magnetic Field Measurements

In this paper we provide a method of handling loop closings in a simultaneous localization and mapping (SLAM) problem by employing indoor magnetic measurements and pose graph optimization. Since the magnetic field in indoor environments has unique spatial features, we can exploit these characteristics to generate the constraints for the pose graph-based SLAM. Specifically, whenever certain motion conditions are satisfied, a series of robot poses along with their magnetic measurements can be grouped into a sequence. A loop closing algorithm is then proposed based on the sequence and applied to the pose graph optimization. Experimental results show that the proposed SLAM system with only wheel encoders and a single magnetometer obtains comparable results with a reference-level SLAM system in terms of robot trajectory, by correctly detecting the loop closings.