Magnetic targets often emit complex magnetic field signals, influenced by both intrinsic and induced magnetic fields. These signals can be difficult to detect and identify due to interference from various types of background and sensor noise. To address this challenge, researchers have developed the ICEEMDAN–RPE–AITD algorithm, promising to improve the clarity of these magnetic signals significantly.
Recent studies have indicated the growing need for effective methods to denoise magnetic field signals, particularly in marine environments where magnetic anomalies can be masked by noise from waves, industrial activities, and solar storms. Recognizing this, the new algorithm incorporates three innovative components: improved complete ensemble empirical mode decomposition with adaptive noise (ICEEMDAN), reverse permutation entropy (RPE), and adaptive interval threshold denoising (AITD).
The ICEEMDAN method works by breaking down the raw magnetic signal data to derive intrinsic mode functions (IMFs), which represent distinct signal components. By calculating the RPE for each IMF, researchers can classify these components based on their signal-to-noise characteristics. This classification results in three categories: signal-dominant IMFs, mixed signal and noise IMFs, and noise-dominant IMFs. The actual signal IMFs are retained, whereas the noisy ones undergo denoising via AITD, helping to reconstruct clear, usable signals.
Initial tests conducted using magnetic field signals from both scaled ship models and real-world measurements validate the effectiveness of the ICEEMDAN–RPE–AITD algorithm. The evaluation measures included the signal-to-noise ratio (SNR) and root mean square error (RMSE), assessing the performance of the denoising process under various conditions.
Notably, the algorithm has proven capable of increasing the SNR significantly—from approximately 5 decibels to over 25 decibels post-denoising—showing up to 137.9% improvement compared to competing methods. This increase not only enhances the reliability of magnetic signal detection but also contributes to more accurate target identification.
One research team stated, "The ICEEMDAN–RPE–AITD exhibits a discernible denoising effect on the ship model magnetic field signals." Such effectiveness is particularly relevant for military and marine operational settings where precise magnetic sensing is imperative.
Among its many advantages, the ICEEMDAN–RPE–AITD algorithm provides remarkable robustness against various noise types, helping it to maintain signal fidelity during turbulent conditions. This algorithm’s development highlights the increasing demand for innovative signal processing technologies capable of thriving under challenging environmental stressors.
From initial assessments, one of the primary conclusions drawn is the vast improvement seen through the application of the ICEEMDAN algorithm framework. It significantly renders magnetic field measurements clearer and smoother, allowing for more accurate interpretations of the data.
Looking forward, the team behind the algorithm plans to explore adaptive optimizations to fine-tune the parameters of the ICEEMDAN method and improve its classifications even more. They are also aiming to tackle non-Gaussian noise issues, potentially broadening the algorithm’s applicability across various domains.
The proposed ICEEMDAN–RPE–AITD algorithm demonstrates promise for the future of magnetic field signal denoising, promising enhanced performance, stability, and effectiveness, especially within complex marine environments.