Analysis of noisy signal restoration quality with Lanczos filter

Abstract

The differential component of the feedback is used to enhance dynamics of controlled mechanisms in control systems. However, the use of the differential component increases the noise interference, resulting in the need to use different filtering methods. This paper describes the research of a differentiated signal input in a differential Lanczos filter. The influence of the filter order on the integrated square error of the filtered signal is defined. The impact of signal dispersion and a sampling interval on the integrated square error of the filtered signal is identified. Moreover, the most convenient parameter values for the Lanczos filter are determined. The paper also includes the comparison of the exponential moving average filter and the Lanczos filter with selected values of noise dispersion and a sampling interval. The research shows the possibility to apply the Lanczos filter for processing of the differentiated signals.