Development of equipment for experimental study of digital algorithms in nonstationary signal processing problems

Abstract

The article deals with the issues of improving the quality of information using ultrasound, as well as existing analog and digital methods of its processing. The characteristics of signal filtering are analyzed depending on the signal-to-noise ratio. A number of modern applications in medical diagnostics are subject of high requirements for accuracy, noise immunity, reliability, continuity of operation and other quality indicators with a high dynamics of changes in non-stationary signals characterizing the functional state of the human body. On this basis, an algorithm and a device capable of extracting the informative components of noisy non-stationary signals are developed. In this paper, an algorithm of processing non-stationary signals of an ultrasonic transceiver is proposed. At the same time, three modules are developed and manufactured: 1) ultrasonic receiver module; 2) ultrasonic noise transmitter module (AM modulation); 3) an ultrasonic signal transmitter module (AM modulation) and an information processing algorithm based on a wavelet-forming function (wavelet threshold) consisting of the Coiflets 5 basis using a heuristically determined threshold value of the wavelet expansion coefficients. In this paper, the wavelet decomposition is carried out up to the 4th level. Based on the analysis of the obtained data, it was concluded that the second decomposition level is the most optimal for filtering non-stationary signals. As the decomposition level increases, the output signal-to-noise ratio decreases, and at the level N = 4, the output signal-to-noise ratio almost does not exceed the input one, therefore, filtering becomes inefficient. As a result of the synthesis of effective parameters of filtering electrocardiosignals, it is found that the maximum degree of signal processing from noise occurs using the Coiflets 5 wavelet using a heuristically determined threshold value. Experimental research is carried out in the training and production laboratory of electronics (FabLab), FSAEI HE “South Ural State University (National Research University)”.