Improved Affine Projection Subband Adaptive Filter for High Background Noise Environments

Haiquan Zhao,Zongsheng Zheng,Zhongyu Wang,Badong Chen
DOI: https://doi.org/10.1016/j.sigpro.2017.02.017
IF: 4.729
2017-01-01
Signal Processing
Abstract:For high background noise environment, an improved affine projection subband adaptive filter (IAPSAF) is proposed, which is derived by minimizing the sum of the squared Euclidean norms of the differences between the updated tap -weight vector and past tap-weight vectors. Due to the use of the past tap-weight vectors, the IAPSAF can achieve lower steady-state misalignment relative to the improved multibandstructured subband adaptive filter (IMSAF). Furthermore, a combined step size IAPSAF (CSS-IAPSAF) is developed to improve the performance of the IAPSAF. Simulation results in an acoustic echo cancellation (AEC) context show that the IAPSAF achieves lower steady-state misalignment than the existing algorithms, such as normalized subband adaptive filter (NSAF), IMSAF, and improved NSAF (INSAF). The improvement of the CSS-IAPSAF is also demonstrated in terms of convergence rate and steady-state misalignment. (C) 2017 Elsevier B.V. All rights reserved.
What problem does this paper attempt to address?