Scholarlink

<p>This work proposes bias compensated sparsity aware linear constrained normalized maximum correntropy criterion (CNMCC) based adaptive filtering algorithms. The proposed algorithms are developed by first integrating ℓ1-norm based sparse penalties into linear constrained normalized maximum correntropy criterion adaptive algorithm and later a bias compensator to reduce the bias generated by input noise. The sparse penalties based on ℓ1-norm result into two algorithms, zero attracted CNMCC (ZA-CNMCC) and reweighted zero attracted CNMCC (RZA-CNMCC). The zero attractor accelerates the convergence of sparse coefficients. Further, the bias compensator is added in ZA-CNMCC and RZA-CNMCC algorithms to mitigate the negative effect of input noise by making estimation unbiased. The usefulness of proposed bias compensated ZA-CNMCC (BC-ZA-CNMCC) and bias compensated RZA-CNMCC (BC-RZA-CNMCC) algorithms are demonstrated by experiments carried out in MATLAB software for linear constrained sparse channel estimation application in the presence of impulsive observation noise against noisy input.</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

Constrained maximum correntropy criterion based sparse algorithm for sparse channel estimation against noisy input

  • 期刊论文

  • 发表日期：

    2025年9月22日

查看全文
  • Rajni Yadav,
  • Kanika Agarwal,
  • Ajay Kumar Gupta

摘要

This work proposes bias compensated sparsity aware linear constrained normalized maximum correntropy criterion (CNMCC) based adaptive filtering algorithms. The proposed algorithms are developed by first integrating ℓ1-norm based sparse penalties into linear constrained normalized maximum correntropy criterion adaptive algorithm and later a bias compensator to reduce the bias generated by input noise. The sparse penalties based on ℓ1-norm result into two algorithms, zero attracted CNMCC (ZA-CNMCC) and reweighted zero attracted CNMCC (RZA-CNMCC). The zero attractor accelerates the convergence of sparse coefficients. Further, the bias compensator is added in ZA-CNMCC and RZA-CNMCC algorithms to mitigate the negative effect of input noise by making estimation unbiased. The usefulness of proposed bias compensated ZA-CNMCC (BC-ZA-CNMCC) and bias compensated RZA-CNMCC (BC-RZA-CNMCC) algorithms are demonstrated by experiments carried out in MATLAB software for linear constrained sparse channel estimation application in the presence of impulsive observation noise against noisy input.