Efficiency of Using Wavelet Transforms for Filtering Noise in the Signals of Measuring Transducers

The article discusses methods of wavelet filtering of noise in signals of measuring transducers using the threshold method of discrete wavelet transform. Special model signals were used to study the methods of wavelet filtering of noise, which enabled estimation of the filtering errors. A method has been developed for determining the parameters with a threshold for all levels of decomposition, which can be used to determine the wavelet function, threshold function, and filtering threshold of the detailing coefficients of the discrete wavelet decomposition. The influence of the parameters of the noise distribution, noise level, number of vanishing moments of the Daubechies wavelet function, nature of the threshold function, and threshold value on the filtering error caused by the noises of non-stationary measuring signals have been investigated by computational experiment. The article presents the results of the study of six threshold functions with the addition of noise to the measuring signal with non-stationary amplitude, frequency, and duty cycle of square-wave pulses. The signal of the Doppler sensors was analyzed and the wavelet filtering parameters were calculated, with the minimum error. The parameters were used to construct graphs of signals before and after filtering directly in the time domain using the inverse wavelet transform.