Proportionate improved normalized subband adaptive filter algorithm for highly noisy sparse system

In order to speed up the convergence of the recently-presented improved normalized subband adaptive filter (INSAF) algorithm in sparse systems, this paper derives its proportionate update based on the concept of the Riemannian distance between the tap-weight vectors. Also, together with the previous improved method of calculating the proportionate matrix, an improved proportionate INSAF (IP-INSAF) algorithm is obtained. Simulation results, in the context of acoustic echo cancellation with a low signal-noise-ratio (SNR), demonstrate the superiority of the proposed algorithm for sparse systems.