Scholarlink

<p>In this paper, we propose two new multichannel adaptive algorithms that have a fast convergence and a low complexity. The proposed algorithms are two types of subband (SB) implementation of the recently published multichannel fast-normalized least mean square (MC-FNLMS) algorithm (i.e Zerouali and Djendi in Phys Commun 62:102233, 2024). The first proposed algorithm is based on the multiband realization of the subband approach, this algorithm has the advantage of high convergence rate. The second algorithm is based on the open loop realization of the subband approach, this later has a lower computational complexity with good convergence rate. A comparative study between the proposed algorithms with the MC-FNLMS, the multichannel Affine Projection Algorithm (MC-APA) and with the classical multichannel subband normalized least mean square (MC-SB-NLMS), have shown performances superiority of the proposed algorithms in terms of convergence rate and computational complexity.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

New Efficient Multichannel Subband Fast Convergence and Low Complexity Algorithms for Adaptive Filtering Applications

  • 期刊论文

  • 发表日期：

    2025年10月5日

查看全文
  • Zerouali Mohamed,
  • Mohamed Djendi

摘要

In this paper, we propose two new multichannel adaptive algorithms that have a fast convergence and a low complexity. The proposed algorithms are two types of subband (SB) implementation of the recently published multichannel fast-normalized least mean square (MC-FNLMS) algorithm (i.e Zerouali and Djendi in Phys Commun 62:102233, 2024). The first proposed algorithm is based on the multiband realization of the subband approach, this algorithm has the advantage of high convergence rate. The second algorithm is based on the open loop realization of the subband approach, this later has a lower computational complexity with good convergence rate. A comparative study between the proposed algorithms with the MC-FNLMS, the multichannel Affine Projection Algorithm (MC-APA) and with the classical multichannel subband normalized least mean square (MC-SB-NLMS), have shown performances superiority of the proposed algorithms in terms of convergence rate and computational complexity.