Wavelet Threshold Denoising Method For Travelling Wave Tube Signal

Noise is one of the important factors that affect the performance of traveling wave tube (TWT). It is difficult to describe the local frequency characteristics of the TWT output signal that contains real signal and non-stationary random noise by using the general signal processing method, so the noise reduction ability is limited. In this paper, we study the threshold-denoising method based on wavelet transform for the demand of the traveling wave tube signal denoising.At first, two modified denoising methods based on single and dual wavelet threshold (SWT-MDM, DWT-MDM) are proposed, respectively. Their mathematical models are established and the relationship between the specific function and the parameters are given. In order to obtain better noise reduction performance, we discuss in detail the effects of decomposition, wavelet, threshold function, threshold selection and other factors on noise reduction, and determine the 5 layers of decomposition with the Haar and Sym6 as the mother wavelets. Three threshold selection algorithms that are suitable for the SWT-MDM are discussed. The threshold selection algorithm for the DWT-MDM is proposed and determined.Secondly, based on the MATLAB simulation platform, the common Blocks, Bumps and Doppler signals as well as the high frequency pulse signal are analyzed by adopting various wavelet threshold-denoising methods, respectively. Simulation results show that the SWT-MDM has better noise reduction performance for the blocks and Bumps signals and DWT-MDM has better effect for Doppler signal and the high frequency pulse signal.Finally, the different types of TWT are measured, and their frequency spectrum distributions of the output signals are obtained. The signals tested are processed in time domain by using SWT-MDM, DWT-MDM and the conventional methods, respectively. The results show that DWT-MDM is better than other methods. Compared with the test results in frequency domain, the signal to noise ratio (SNR) for the tube 1 is improved by 5.3245dB and SNR for the tube 2 is improved by 5.2496dB, both after are denoised.