Magnetic Anomaly-Matched Trajectory and Dead Reckoning Fusion Mobile Robot Navigation

Abstract

Environments with varying magnetic field distortion cannot be navigated stably with magnetic anomaly based navigation algorithms.In this study, we propose a stable navigation solution for various indoor environments by fusing magnetic anomaly matched trajectories and mobile robot inertial trajectories.The proposed method uses dead reckoning as the primary navigation system and compensates for the navigation sensor error with a feedback structure through the optimization of the anomaly matching trajectory and dead reckoning trajectory.In addition, by determining the trajectory key-frame, the extended Kalman filter measurement update is performed using only the localization results with high accuracy.An open dataset was used to verify the performance of the algorithm, which was compared with existing algorithms.The proposed method is cost-effective, because the proposed method uses only an odometer, gyroscope, and magnetometer for indoor navigation.