Magnetic Interference Compensation Method for a Deep-Sea Human-Occupied Vehicle Based on Dynamic Excitation

Human-occupied vehicles (HOVs) provide an ideal platform for high-resolution near-bottom magnetic anomaly detection. However, complex platform-generated magnetic interference severely limits the reliable extraction of weak magnetic signals. The conventional Tolles–Lawson (T–L) model relies on large-amplitude attitude maneuvers to estimate interference coefficients, but such maneuvers are infeasible for HOVs because of their large inertia, hydrodynamic coupling, and deep-sea safety constraints. To address this limitation, we propose a dynamic-excitation magnetic interference compensation method that exploits the inherent dynamic characteristics of the platform. By commanding the HOV to execute acceleration–deceleration cycles and horizontal S-shaped turns, the method indirectly excites pitch and roll variations through coupled vehicle dynamics and provides the attitude-dependent information required by the complete 18-term T–L model. The 18-term basis set is constructed from measured direction cosines and their time derivatives, with coefficient identifiability determined by the excitation data. Helicopter-based analog experiments showed that the proposed maneuver yielded improvement ratios of 6.6 and 12.1 on two test lines, comparable to the values of 6.3 and 12.5 obtained using conventional airborne calibration maneuvers. In an in situ trial with the Shenhai Yongshi (“Deep-Sea Warrior”) HOV, the method reduced the magnetic-field standard deviation from 0.8870 nT to 0.4157 nT, corresponding to an improvement ratio of 2.13. The compensated record exhibited a residual dynamic-field variation of 0.4157 nT under the tested maneuvering conditions. The helicopter experiment evaluated the shared excitation sequence and signal-processing workflow on a controllable airborne platform. These results demonstrate a practical engineering approach to suppressing dynamic magnetic-field variations below 1 nT in the tested HOV conditions.