Combined distributed incremental affine projection algorithm for acoustic echo cancellation

To improve the performance of the conventional distributed incremental affine projection algorithm (DIAPA) in acoustic echo cancellation, a combined distributed incremental affine projection algorithm is proposed in this brief. The proposed algorithm combines two DIAPA equipped with different projection orders by employing a mixing scalar parameter, where one with high projection order guarantees fast convergence rate for the initial stage and the other with low projection order maintains small misalignment for the steady stage. The mixing scalar parameter is endowed with capability to adapt itself to be close to one in the initial stage and approximate to zero in the steady-state stage. Moreover, the analysis of the mixing scalar parameter is presented. We also take into account the computational complexity of the proposed algorithm. The simulations conducted in the system identification and acoustic echo cancellation experiments illustrate the superiority of our findings. © 2018, Springer Science+Business Media, LLC, part of Springer Nature.