Unbiased equation-error adaptive IIR filtering based on monicnormalization

摘要

We present a novel may to remove the bias in equation-error basednadaptive infinite impulse response (IIR) filtering by conceiving anscheme called monic normalization. It is found that normalizing all thencoefficients of the denominator filter by the first coefficient afterneach adaptation removes the bias and leads to unbiased estimates. Thenanalysis of stationary points is presented to show that the proposednmethod can indeed produce unbiased parameter estimates in the presencenof noise. The computer simulation results also demonstrate that thenproposed method performs better than or comparable to existingnalgorithms, while requiring much lower computational complexity