Interference Suppression for Geomagnetic Sensors Based on Empirical Modal Decomposition and Feature Clustering

To address the issue of heading deviation in geomagnetic navigation caused by environmental noise and pulse interference from the motor drive system, this paper proposes an interference suppression method for geomagnetic signals based on Hilbert-Huang Transform (HHT) and feature clustering analysis. To avoid the frequency aliasing and distortion caused by empirical mode decomposition (EMD), the raw signal is first decomposed using EMD. Through feature clustering analysis, the intrinsic mode functions (IMFs) obtained from the decomposition are classified into noise IMFs and mixed IMFs based on the frequency similarity criterion with geomagnetic variations. Then, an adaptive least mean square (LMS) filtering algorithm is proposed for denoising the mixed IMFs. This algorithm can adaptively adjust the filtering parameters according to the noise level in the mixed IMFs. Finally, the denoised mixed IMFs and the residuals are reconstructed to obtain the final geomagnetic signal. Experimental results show that compared with traditional MF, IMMF, and MF-EMD methods, the proposed improved empirical model decomposition with feature clustering (FC-IEMD) method can effectively enhance the signal-to-noise ratio (SNR) while reducing the root mean square error (RMSE).