Scholarlink

<p>This paper investigates the advantages of learning-based approaches over classical methods and introduces a novel data-driven framework for Direction of Arrival (DOA) estimation. The proposed neural network architecture uniquely integrates a modified deep unfolding block in a well-established structure, tailored for resource-constrained one-bit hardware environments. The study also evaluates these approaches under mutual coupling, a prevalent real-world impairment. Focusing on one-bit quantization, the proposed method addresses the need for reduced system complexity and power consumption, providing clear benefits for power-efficient, cost-effective, and hardware-limited applications such as massive multiple-input multiple-output (MIMO) systems and automotive radars. The performance of the proposed approach is compared against both conventional and robust state-of-the-art algorithms. Simulation results demonstrate enhanced robustness at various signal-to-noise ratios under both ideal and mutual coupled conditions.</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

Learning-Based One-Bit Direction of Arrival Estimation with Mutual Coupling

  • 期刊论文

  • 发表日期：

    2026年5月16日

查看全文
  • Yasin Azhdari,
  • Mahmoud Farhang

摘要

This paper investigates the advantages of learning-based approaches over classical methods and introduces a novel data-driven framework for Direction of Arrival (DOA) estimation. The proposed neural network architecture uniquely integrates a modified deep unfolding block in a well-established structure, tailored for resource-constrained one-bit hardware environments. The study also evaluates these approaches under mutual coupling, a prevalent real-world impairment. Focusing on one-bit quantization, the proposed method addresses the need for reduced system complexity and power consumption, providing clear benefits for power-efficient, cost-effective, and hardware-limited applications such as massive multiple-input multiple-output (MIMO) systems and automotive radars. The performance of the proposed approach is compared against both conventional and robust state-of-the-art algorithms. Simulation results demonstrate enhanced robustness at various signal-to-noise ratios under both ideal and mutual coupled conditions.