Set-membership versions of improved normalized subband adaptive filter algorithm for highly noisy system

In order to improve the performances of the recently-presented improved normalized subband adaptive filter (INSAF) algorithm for highly noisy system, this paper proposes a set-membership version of the INSAF algorithm (SM-INSAF) by exploiting the concept of the set-membership filtering. Apart from obtaining lower steady-state error under the same convergence rate, the proposed algorithm significantly reduces the overall computational complexity. In addition, to further reduce the steady-state error of the SM-INSAF, its smooth variant is developed by using smooth subband output errors to update the step sizes, called the SSM-INSAF algorithm. Simulation results in low signal-noise-ratio (SNR) environments, demonstrate the superiority of the proposed algorithms.