A Shrinkage Dual Sign Subband Adaptive Filter with Individual Weight Factor

A dual sign subband adaptive filter with individual weight factor (IWF-DuSSAF) is proposed, which can increase the convergence rate with a little additional computational burden compared with the sign subband adaptive filter with individual weight factor (IWF-SSAF). The dual sign method is used for deriving the proposed IWF-DuSSAF algorithm. Selecting the proper parameters of proposed IWF-DuSSAF algorithm will lead to small steady-state errors and fast convergence. The proposed sign subband adaptive filter with individual weight factor has the tradeoff between steady-state misalignment and convergence rate. To overcome this drawback, we present the shrinkage variable step-size IWF-DuSSAF (SHIWF-DuSSAF) algorithm. To make the proposed SHIWF-DuSSAF algorithm robust against impulsive noise, a novel robust shrinkage method is adopted to estimate the noise free error signal, which uses two threshold parameters. We gain the threshold by deciding the confidence degree. The occurrence of impulsive noise is detected by novel method. Finally, simulations are carried out to demonstrate the advantages of the proposed IWF-DuSSAF and SHIWF-DuSSAF algorithms.