基于小波能量比和改进阈值函数的通信信号降噪算法

打开文本图片集

中图分类号：TN911.3 文献标志码：A DOI：10.12305/j.issn.1001-506X.2025.07.03

Abstract：To improve thequality of communication signals with low signal-to-noise ratio （SNR），limited samplingrates，and unknown center frequencies，and to enhance recognition performance，this paper implements adaptive estimation of the center frequency and proposes an improved wavelet denoising algorithm. The central frequency estimation section realizes rough clasification based on the differences in the power spectrum of 11 types of communication signals，and improves the frequency centering method based on diferent clasification results toachieve the estimation of the central frequency.The improved wavelet denoising algorithm addreses issues with soft and hard threshold functions by proposing a parameter-adjustable and continuous wavelet threshold function. Additionally，it uses wavelet energy ratios to characterize the energy distribution of wavelet coeficients for communication signals with diffrent center frequencies，applying different processing methods based on their magnitudes.Finall，modulation recognition experiments are conducted on 1l types of communication signals within an SNR range of [-10，10] dB. Simulation results show that the proposed denoising algorithm achieves notable noise reduction for all11 types of communication signals，exhibiting an improvement of 10%-40% in the average signal recognition rate within the SNR range of [-10，0] dB compared to the unprocessed signals.

Keywords： low signal-to-noise ratio（SNR）； unknown center frequency； improved wavelet threshold; wavelet energy ratio；modulation recognition

0 引言

调制识别作为信号处理领域的一个重要分支，在军事通信[1-2]、民用通信[3]等诸多领域中都发挥着不可或缺的作用。（剩余19122字）