Abstract:This study proposes a sensor integration design of magnetic positioning system (MPS) for achieving robust navigation performance. Despite consistent permeability of AC magnetic fields, signal strength that is inversely proportional to the square of the distance greatly constrains the operational range of the MPS. Additionally, the signal becomes vulnerable to environmental noise with deteriorated signal-to-noise ratio, which may bring about partial signal outage from some magnetic transmitter. In this background, this paper proposes the tightly coupled MPS/INS integrated navigation system to overcome these shortcomings. Instead of implementing isolated pose estimation via multiple magnetic measurements, it is designed each magnetic field vector is independently incorporated in the INS mechanization framework. As a result, successful filter correction has been guaranteed in the constrained magnetic measurement environments, thereby obtaining results that ensure pose estimation performance in an extended range with poor signal quality or even in a partial transmitter outage case. Besides, the proposed method contributes to demonstrate improved estimation performance in the navigation filter by providing more accurate magnetic predictions and measurement residuals compared to the traditional dipole technique. The proposed algorithm is demonstrated through trajectory experiments with an unmanned vehicle platform in outdoor domain with different ranges and partial signal occlusion scenario. When compared with the reference system, the robust navigation results using the MPS/INS TC algorithm exhibited pose estimation error of decimeters in position and 2 degrees in attitude along with computational efficiency. Moreover, robust estimation is observed even in a partial signal occlusion in contrast with previous diverging results from MPS-only or MPS/INS loosely coupled system.

Design and Performance Validation of Tightly Coupled Magneto-Inertial Navigation System for Robust Outdoor Application