Scholarlink

<p>This paper addresses the problem of nonlinear system identification under unknown but bounded noise. A kernel-based set-membership identification method is proposed, which extends the ellipsoidal outer bounding (EOB) algorithm to nonlinear systems through kernel learning techniques. The proposed approach maps the input data into a reproducing kernel Hilbert space (RKHS), enabling the nonlinear identification problem to be reformulated as a linear regression model in the feature space. The proposed algorithm recursively computes parameter estimates for nonlinear systems. A convergence analysis is established, and the effectiveness of the proposed method is demonstrated through numerical simulations on two Non-Line-of-Sight (NLOS) outdoor and indoor channel identification case studies.</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

Kernel-Based Machine Learning Ellipsoidal Outer Bounding for Non-Line-of-Sight Outdoor and Indoor Channel Identification

  • 期刊论文

  • 发表日期：

    2026年2月22日

查看全文
  • Hasna El Maizi,
  • Rachid Fateh,
  • Mathieu Pouliquen,
  • Said Safi,
  • Miloud Frikel

摘要

This paper addresses the problem of nonlinear system identification under unknown but bounded noise. A kernel-based set-membership identification method is proposed, which extends the ellipsoidal outer bounding (EOB) algorithm to nonlinear systems through kernel learning techniques. The proposed approach maps the input data into a reproducing kernel Hilbert space (RKHS), enabling the nonlinear identification problem to be reformulated as a linear regression model in the feature space. The proposed algorithm recursively computes parameter estimates for nonlinear systems. A convergence analysis is established, and the effectiveness of the proposed method is demonstrated through numerical simulations on two Non-Line-of-Sight (NLOS) outdoor and indoor channel identification case studies.