Scholarlink

<p>The complex correntropy has a wide range of applications for adaptive filtering in the complex domain due to its effectiveness in measuring local similarity of complex variables. To improve the filtering performance of complex adaptive filtering algorithms when dealing with the contaminated input and output signals, this paper focuses on the complex-valued error-in-variables (CEIV) model in the presence of non-Gaussian noise. First, a maximum mixture total complex correntropy (MMTCC) algorithm for the CEIV model is proposed by utilizing the mixture correntropy (MC) criterion and total least squares (TLS) method. Then, a variable mixture coefficient strategy is presented to avoid the selection of the mixture coefficient, generating the variable MMTCC (VMMTCC) algorithm. Moreover, the local stability analysis, steady-state performance analysis, and convergence condition of the MMTCC algorithm are provided for theoretical analysis. Finally, simulations verify the correctness of the theoretical analysis and the superiorities of the proposed MMTCC and VMMTCC algorithms in the CEIV model.</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

Maximum mixture total complex correntropy for adaptive filtering

  • 期刊论文

  • 发表日期：

    2025年1月17日

查看全文
  • Peng Cai,
  • Yiji Chen,
  • Dongyuan Lin,
  • Yunfei Zheng,
  • Zhongyuan Guo,
  • Shiyuan Wang

摘要

The complex correntropy has a wide range of applications for adaptive filtering in the complex domain due to its effectiveness in measuring local similarity of complex variables. To improve the filtering performance of complex adaptive filtering algorithms when dealing with the contaminated input and output signals, this paper focuses on the complex-valued error-in-variables (CEIV) model in the presence of non-Gaussian noise. First, a maximum mixture total complex correntropy (MMTCC) algorithm for the CEIV model is proposed by utilizing the mixture correntropy (MC) criterion and total least squares (TLS) method. Then, a variable mixture coefficient strategy is presented to avoid the selection of the mixture coefficient, generating the variable MMTCC (VMMTCC) algorithm. Moreover, the local stability analysis, steady-state performance analysis, and convergence condition of the MMTCC algorithm are provided for theoretical analysis. Finally, simulations verify the correctness of the theoretical analysis and the superiorities of the proposed MMTCC and VMMTCC algorithms in the CEIV model.