An effective LFM signal reconstruction method for signal denoising

Abstract

One of the main challenges in signal denoising is to accurately restore useful signals in low signal-to-noise ratio (SNR) scenarios. In this paper, we investigate the signal denoising problem for multi-component linear frequency modulated (LFM) signals. An effective time-frequency (TF) analysis-based approach is proposed. Compared to the existing approaches, our proposed one can further increase the noise suppressing performance and improve the quality of the reconstructed signal. Experimental results are presented to show that the proposed denoising approach is able to effectively separate the multi-component LFM signal from the strong noise environments.