Determining conditions preserving information in smothing of a spectrum of noisy electro cardio graphy.

Abstract

Information of Bio-signals is often corrupted by additive noise. Mono-frequency characterization of such noisy with random translation is recognized as major obstacle for denoising. The usual method of denoising is to pass the composite noisy signal throught an adaptive notch filter (ANF), which originated with the pioneering work of B. Widrow. An ANF has only the ability to adjust their parameters automatically with slow adaptation. This was a major motivation for enhancing the convergence rate of processing with the accuracy of the denoising. Our proposed method should have an optimum speed of adaptation and allow the minimization of loss of information of the signal of interest [6]. The results presented in [6] suggest the need to minimize a error of  smoothing of a spectrum of the noisy signal and to solve problems of determining the affect on upper bound of probability of event error threshold exceeded. This paper deals with new method for such problem and its basis is stated into a lemma. In addition, extensive simulations are conducted to corroborate the efficiency of the proposed estimator with the theoretical analysis.