Steady-state comparative performance evaluation of piloted adaptive notch filter

Abstract

In this paper, we have presented the performance comparison of the piloted adaptive notch filter with the second- order adaptive IIR lattice notch filter and the adaptive IIR notch filter with constrained poles and zeroes. The steady-state analysis for mean square error (MSE) of the frequency estimate using simulations has been provided with roles of the user variables such as notch bandwidth and signal to noise ratio (SNR). This study verifies the excellent performance exhibited by the piloted adaptive notch filter over other algorithms with respect to the speed of convergence and MSE. Extensive computer simulations have been done to study the performance over a wide range of SNR and frequency values. I. INTRODUCTION In digital signal processing, frequency estimation and track- ing of sinusoids in noise has been an important area of research to solve problems in radar, communications, bio-medical and other related areas. Extensive research has been carried out in designing and analysis of both finite-impulse response (FIR) type adaptive filters and infinite-impulse response (IIR) type adaptive filters to estimate and track the frequency of input sinusoid being corrupted with noise. The FIR based adaptive line enhancer (ALE), though the simplest solution to the problem, has relatively slow rate of convergence and is highly responsive to the autocorrelation of the input signal. It requires a relatively larger number of filter coefficients as compared to the IIR based ALE for the same narrow bandwidth notch to reject the interfering noise. Thus, there have been efforts to improve the performance of adaptive IIR notch filters and various novel filter structures have been proposed in literature.