Bias-Compensated Sign Subband Adaptive Filter Algorithm With Individual Weighting Factors for Input Noise.

To thoroughly exploit the decorrelation property of the traditional sign subband adaptive filter (SSAF), an individual weighting factors SSAF (IWF-SSAF) was presented. However, when the input of the adaptive filter is polluted by white Gaussian noise, the IWF-SSAF will produce estimation bias for unknown system identification. Thus, this brief proposes the bias-compensated IWF-SSAF (BC-IWF-SSAF) for resolving the above trouble. Especially, the unbiased criterion is applied to obtain compensation term for the conventional IWF-SSAF to alleviate the effect of noisy input. The proposed BC-IWF-SSAF both retains stability of impulse noise and decreases the estimated bias. The stable condition of the proposed BC-IWF-SSAF is achieved resorting to Price’s theorem with some reasonable assumptions and theorems. In addition, the complexity is given as well. Finally, simulations confirm the superiority of the proposed BC-IWF-SSAF algorithm. [ABSTRACT FROM AUTHOR]