Signal noise reduction method based on parametric optimization of variational mode decomposition

To address the issue of significant noise and interference in satellite communication channels, this study introduces a noise reduction approach that employs Variational Mode Decomposition (VMD) enhanced by adaptive parameter optimization. Given the substantial impact of the number of decompositions and the penalty factors on the outcomes of VMD, this research develops a strategy for parameter optimization. Initially, it utilizes the count of effective spectral peaks in the singular value difference spectrum to ascertain the signal count in scenarios of multiple signal overlays, subsequently optimizing the modal layer count for VMD decomposition. Furthermore, it refines the penalty factor by comparing the difference in central frequency between adjacent modes against a set threshold. The testing of actual signals demonstrates that the parameter optimization strategy not only addresses the challenge of determining decomposition parameters in the traditional VMD algorithm but also mitigates the potential for information loss or over-decomposition. Moreover, this method demonstrates effective noise reduction in channel environments characterized by intense noise and interference, offering valuable insights for channel noise mitigation.