| Title |
Sparsity-aware Affine Projection Least Mean Mixed-norm Algorithm for Sparse System |
| Authors |
이민호(Minho Lee) ; 조태성(Taesung Cho) ; 박부견(PooGyeon Park) |
| DOI |
https://doi.org/10.5573/ieie.2022.59.5.91 |
| Keywords |
Adaptive filter; Affine projection least mean mixed-norm (AP-LMMN) algorithm; mixed-norm; sparse system; zero attraction |
| Abstract |
In this paper, we propose sparsity-aware affine projection least mean mixed-norm (SAP-LMMN) algorithms for system identification of the sparse systems used in adaptive filters. The proposed algorithms are derived from minimizing the mixed-norm cost function of and norms with the sparsity penalty. We introduce the norm or norm penalties and compare the performance of both algorithms. The zero-attracting terms derived from the norm and norm promote sparsity in the weight of the adaptive filter and accelerate convergence during the weight update process. Simulations are conducted in the sparse system identification scenarios to verify the proposed algorithms. Through acoustic echo cancellation system simulation, it is confirmed that the proposed algorithm has better performance than conventional algorithms. |