Stereophonic Acoustic Echo Cancellation Using an Improved Affine Projection Algorithm with Adaptive Multiple Sub-Filters

Stereophonic acoustic echo cancellation faces major challenges due to the high correlation and non-circular nature of stereo input signals. It causes lower echo cancellation performance and reduced efficacy of conventional adaptive echo cancellation algorithms. To mitigate these research gaps, this paper proposes a multiple sub-filters-based variable step-size improved affine projection algorithm to ensure better convergence and faster adaptability for dynamic stereophonic echo paths. In the proposed system, a recursive mean square deviation has been utilized for dynamic adaptation, and for rapid tracking of dynamic changes in the acoustic echo path, a scheme to reset the step-size has also been proposed. The simulation results for both static and dynamic conditions show that the proposed algorithm achieves improved echo cancellation and faster convergence in real-time situations.