Identification of dead time using wavelet

Abstract

In this paper, the authors discuss a method of identification of the dead time in a linear system using a wavelet. The authors use nonstationary step-like functions as the test signal. First a set of differential operators are proposed for calculating the differentials of an observed signal by the Daubechies wavelet. The differential operators are constructed by iterative projections of the differential of the scaling function for a multiresolution analysis into a dilation subspace. Using the above differentials, they propose a method to estimate the parameters of a linear system for a priori assumed dead time. In order to estimate the dead time of a transfer function they evaluate a criterion function for several values of the dead time. The value that maximizes the criterion function becomes the estimate of the dead time. They present several numerical examples. It is shown that the proposed method is effective in identification of dead time.