Scholarlink

<p>The integrated sensing and communication (ISAC) in 6G terahertz (THz) ultra-massive multiple-input multiple-output (UM-MIMO) systems faces challenges in near-field scenarios, where spherical wavefronts introduce coupled angle, distance. Existing schemes often address localization and CSI estimation separately, neglecting their interdependence. This paper proposes a joint uplink localization and CSI estimation scheme for hybrid analog-digital THz UM-MIMO systems. First, a coarse angle-of-arrival (AoA) estimation method is developed using DFT-based analog combiners, exploiting near-field angular spread effects. Then, a refined localization scheme iteratively optimizes angle and distance parameters via dynamic interval contraction, reducing computational complexity. Further, a line-of-sight (LoS)-prior-enhanced CSI estimation (LPE-CE) method decouples LoS and non-line-of-sight (NLoS) components using subspace-orthogonal combiners and introduces an adaptive polar-domain codebook that dynamically expands based on residual thresholds. Simulations demonstrate the proposed scheme achieves sub-meter localization accuracy with low complexity, while LPE-CE outperforms benchmarks and adaptively reduces codebook size efficiently.</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

Joint localization and channel estimation for terahertz near-field ISAC UM-MIMO systems

  • 期刊论文

  • 发表日期：

    2025年11月18日

查看全文
  • Yanran Sun,
  • Chuang Yang,
  • Yuheng Fan,
  • Renzhi Yuan,
  • Mugen Peng

摘要

The integrated sensing and communication (ISAC) in 6G terahertz (THz) ultra-massive multiple-input multiple-output (UM-MIMO) systems faces challenges in near-field scenarios, where spherical wavefronts introduce coupled angle, distance. Existing schemes often address localization and CSI estimation separately, neglecting their interdependence. This paper proposes a joint uplink localization and CSI estimation scheme for hybrid analog-digital THz UM-MIMO systems. First, a coarse angle-of-arrival (AoA) estimation method is developed using DFT-based analog combiners, exploiting near-field angular spread effects. Then, a refined localization scheme iteratively optimizes angle and distance parameters via dynamic interval contraction, reducing computational complexity. Further, a line-of-sight (LoS)-prior-enhanced CSI estimation (LPE-CE) method decouples LoS and non-line-of-sight (NLoS) components using subspace-orthogonal combiners and introduces an adaptive polar-domain codebook that dynamically expands based on residual thresholds. Simulations demonstrate the proposed scheme achieves sub-meter localization accuracy with low complexity, while LPE-CE outperforms benchmarks and adaptively reduces codebook size efficiently.