Enhancement and Tracking of a Single Sinusoid in Noise Using Cumulant-Based IIR Adaptive Notch Filter

概要

論文の詳細を見る
In this paper, a novel cumulant-based adaptive notch filtering technique for the enhancement and tracking of a single sinusoid in additive noise is presented. In this technique, the enhanced signal is obtained as the output of a narrow band-pass filter implemented using a second-order pole-zero constraint IIR adaptive notch filter, which needs only one coefficient to be updated. The filter coefficient, which leads to identifying and tracking the sinusoidal frequency, is updated using a suggested adaptive algorithm employing a recursive estimate of the kurtosis and only one-sample-lag point of a selected one-dimensional fourth-order cumulant slice of the input signal. Therefore, the proposed technique provides automatically resistance to additive Gaussian noise. It is also shown that the presented technique outperforms the correlation-based counterpart in handling additive non-Gaussian noise. Simulation results are provided to show the effectiveness of the proposed algorithm in comparison with the correlation-based lattice algorithm.
社団法人電子情報通信学会の論文
2001-02-01