Scholarlink

In the present study, we propose new implementation of four variable step-size estimation methods on the -law Proportionate Normalized Least-Means-Square (MPNLMS) algorithm for acoustic system identification and acoustic echo cancellation (AEC). These four methods aim to enhance the convergence speed and achieve low steady state misadjustment. The obtained results demonstrate that the four variables step-sizes MPNLMS (VSS-MPNLMS) algorithms are more effective compared to their fixed step-size.

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

Efficient Variable μ_law Proportionate Algorithms for Acoustic Echo Cancellation and System Identification

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Ali Alilouche,
  • Rédha Bendoumia,
  • Abderezzak Guessoum,
  • Islam Hassani

摘要

In the present study, we propose new implementation of four variable step-size estimation methods on the -law Proportionate Normalized Least-Means-Square (MPNLMS) algorithm for acoustic system identification and acoustic echo cancellation (AEC). These four methods aim to enhance the convergence speed and achieve low steady state misadjustment. The obtained results demonstrate that the four variables step-sizes MPNLMS (VSS-MPNLMS) algorithms are more effective compared to their fixed step-size.