Research Of Denoising Algorithm For Radio Frequency Signal

Nowadays,the application fields of radio frequency signals are becoming more and more extensive,such as: wireless communication,RFID radio frequency detection and identification systems,radar communication systems,etc.RF signals are high-frequency signals that are modulated to facilitate channel transmission.However,the inevitability of noise interference will cause errors in the information analysis of the received signal at the receiving end,which will have a poor effect on the subsequent research work,so the significance of denoising is self-evident.At present,there are many denoising algorithms in the field of signal processing.They all have advantages,disadvantages and application conditions,but most of them are applied to images and low-frequency signals,and there are few researches on RF signals relatively,which makes the subject of RF signal denoising algorithm has important value and significance.This paper focuses on wavelet analysis theory and singular value decomposition theory,studies their feasibility of denoising,and applies them to RF signals.The main tasks are as follows:First,the background and significance of RF signal denoising are summarized,the status and progress of research at home and abroad are briefly described,and the common noise in communication systems,RF noise signal models and common denoising indicators are introduced.Secondly,the relebant theories of wavelet analysis are introduced in detail,including:multi-resolution analysis,continuous and discrete wavelet transform,Mallat fast algorithm,wavelet basis mathematical characteristics and selection principles,and commonly used wavelet functions.The wavelet threshold denoising algorithm is applied to RF signals,the threshold application method and estimation criteria are introduced,and the experiment simulation and comparative analysis are carried out.The results show that: for the simulation signals in this chapter,different wavelet parameters are selected,and the direct zeroing of high-frequency coefficients is invalid.The method using soft,hard threshold functions and multiple threshold estimation criteria has achieved good denoising effects.The signal-to-noise ratio was increased about 4.9～7.8d B respectively,and the root mean square error and cross-correlation coefficient indexes were also improved.Finally,the singular value decomposition denoising theory is introduced,the influence of the structure and dimension of the decomposition matrix and singular value processing on the denoising results are analyzed through simulation,and a joint denoising algorithm of wavelet and SVD in two combined ways is proposed.The results show that:for the simulation signal in this chapter,the Hankel matrix is used to construct the signal,and the obtained denoising effect is good,and when the matrix dimensions are similar and the first two main singular values are selected to reconstruct the signal,the result is the best,and the SNR ratio is improved about 30.4d B.At the same time,the denoising effect of the second combination of wavelet and SVD is better than that of the first,which improves the SNR ratio of the wavelet high-frequency coefficient zeroing method by about 18.0d B.