Scholarlink

<p>This paper presents a Dynamic Dual-Adaptation Filter (DDAF) aimed at improving the robustness and convergence of adaptive filters operating in impulsive, non-Gaussian noise environments. Traditional adaptive algorithms such as Least Mean Squares (LMS) and Normalized LMS (NLMS) assume Gaussian noise and tend to perform poorly when the signal is corrupted by heavy-tailed or impulsive disturbances. To address this limitation, the proposed DDAF employs a dual adaptation mechanism in which both the filter step-size (<InlineEquation ID="IEq1"> <EquationSource Format="TEX">\(\:\mu\:\)</EquationSource> </InlineEquation>) and the nonlinearity parameter (<InlineEquation ID="IEq2"> <EquationSource Format="TEX">\(\:p\)</EquationSource> </InlineEquation>) are updated simultaneously on a sample-by-sample basis. The adaptive step-size is derived from an energy-based stability criterion, while the nonlinearity parameter is dynamically tuned according to the projection error magnitude. This strategy allows the algorithm to automatically adjust between fast convergence and strong noise immunity without requiring prior parameter tuning. Simulation studies conducted under varying impulsive noise conditions demonstrate that the proposed approach provides improved steady-state performance and higher signal-to-noise ratio compared to conventional LMS, NLMS, and Fixed Least-Mean p-Power (FxLMP) algorithms.</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

Robust adaptive filtering through dynamic dual-adaptation in α-stable conditions

  • 期刊论文

  • 发表日期：

    2025年12月27日

查看全文
  • V. Saravanan,
  • N. Santhiyakumari,
  • R. Hemalatha,
  • P. Shanmuga Sundaram

摘要

This paper presents a Dynamic Dual-Adaptation Filter (DDAF) aimed at improving the robustness and convergence of adaptive filters operating in impulsive, non-Gaussian noise environments. Traditional adaptive algorithms such as Least Mean Squares (LMS) and Normalized LMS (NLMS) assume Gaussian noise and tend to perform poorly when the signal is corrupted by heavy-tailed or impulsive disturbances. To address this limitation, the proposed DDAF employs a dual adaptation mechanism in which both the filter step-size ( \(\:\mu\:\) ) and the nonlinearity parameter ( \(\:p\) ) are updated simultaneously on a sample-by-sample basis. The adaptive step-size is derived from an energy-based stability criterion, while the nonlinearity parameter is dynamically tuned according to the projection error magnitude. This strategy allows the algorithm to automatically adjust between fast convergence and strong noise immunity without requiring prior parameter tuning. Simulation studies conducted under varying impulsive noise conditions demonstrate that the proposed approach provides improved steady-state performance and higher signal-to-noise ratio compared to conventional LMS, NLMS, and Fixed Least-Mean p-Power (FxLMP) algorithms.