Physics-Based Modeling and Compensation of Magnetic Interference From Permanent-Magnet Synchronous Motors

Accurate magnetic field measurement is crucial for navigation and state estimation in tightly integrated unmanned systems. However, permanent magnet synchronous motors (PMSMs) introduce strong and spatially varying near-field interference that cannot be effectively compensated by conventional platform-level disturbance models. This article develops a physics-informed and parameterizable interference model that explicitly incorporates the stator current excitation field, the rotor magnetic dipole rotational field, and the gearbox-induced magnetic distortion field, while introducing a frequency-dependent shielding factor to characterize eddy-current attenuation caused by the motor housing and structural materials. To enable real-time implementation, the shielding term is approximated via a first-order Taylor expansion, yielding a linearly parameterized formulation suitable for recursive least squares (RLSs) identification without compromising physical interpretability. Experimental validation on both a fixed single-motor testbench and a quadruped robot equipped with multiple PMSMs demonstrates that the proposed method reduces interference standard deviation from 430.34 to 28.95 nT in the single-motor case and from 586.01 to 60.68 nT under multimotor coupled operation. These results demonstrate the effectiveness and practical applicability of the proposed approach under the tested conditions, providing a theoretically grounded and practically feasible solution for high-precision magnetic measurement in compact multimotor robotic platforms.