An implementation of high-speed adaptive noise canceller with parallel block structure

Abstract

An adaptive algorithm for detecting a single sinusoid of unknown frequency corrupted by Gaussian noise has been proposed by Nishimura et al. using IIR bandpass filter with a variable center frequency/sup [1]/. The filter has the parallel block structure for fast processing. However, the algorithm has two problems: one is that it has several input frequencies making it impossible to converge; another is that the convergence rate cannot be higher than that of the scalar structure. In this paper, the causes of these problems are investigated and a new algorithm is proposed to solve these problems. Simulation results are given to illustrate the performance of the proposed algorithm.