Scholarlink

<p>Integrated sensing and communication (ISAC) is a key enabler for next-generation wireless networks, supporting joint use of spectrum and hardware for both sensing and communication. Reconfigurable intelligent surfaces (RIS), including simultaneously transmitting and reflecting (STAR)-RIS, improve flexibility by dynamically shaping the wireless environment. Fully leveraging RIS-assisted ISAC requires accurate channel estimation, particularly in scenarios with multiple targets, users, and diverse spatial configurations. This work proposes a unified learning-based channel estimation framework for multi-target, multi-RIS-assisted ISAC. For the direct sensing link, a parametric learning model captures the physical structure and incorporates a residual refinement path for robustness. For communication links, two end-to-end denoising architectures are proposed: one with a fully connected subnetwork and another using a fixed extreme learning machine layer for lightweight user-side processing. Simulations under realistic conditions demonstrate strong estimation accuracy and generalization, achieving up to 94% complexity reduction over existing methods, enabling efficient deployment of RIS-assisted ISAC systems.</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

Hybrid angle-parametric and denoising-based channel estimation for multi-RIS multi-target ISAC systems

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年6月3日

查看全文
  • Alice Faisal,
  • Ibrahim Al-Nahhal,
  • Octavia A. Dobre

摘要

Integrated sensing and communication (ISAC) is a key enabler for next-generation wireless networks, supporting joint use of spectrum and hardware for both sensing and communication. Reconfigurable intelligent surfaces (RIS), including simultaneously transmitting and reflecting (STAR)-RIS, improve flexibility by dynamically shaping the wireless environment. Fully leveraging RIS-assisted ISAC requires accurate channel estimation, particularly in scenarios with multiple targets, users, and diverse spatial configurations. This work proposes a unified learning-based channel estimation framework for multi-target, multi-RIS-assisted ISAC. For the direct sensing link, a parametric learning model captures the physical structure and incorporates a residual refinement path for robustness. For communication links, two end-to-end denoising architectures are proposed: one with a fully connected subnetwork and another using a fixed extreme learning machine layer for lightweight user-side processing. Simulations under realistic conditions demonstrate strong estimation accuracy and generalization, achieving up to 94% complexity reduction over existing methods, enabling efficient deployment of RIS-assisted ISAC systems.