Scholarlink

<p>Accurate localization and energy-efficient clustering are critical for Wireless Sensor Networks (WSNs), especially in dynamic and resource-constrained environments. Traditional methods struggle to balance precision, computational cost, and energy consumption, often leading to reduced scalability and network lifetime. This paper presents a novel Label-aware Gates-Controlled Deep Unfolding Network with Brown Bear Optimization (LGDU-BBO) for high-precision WSN localization. The framework integrates a Label-aware Attention Network (L2AN) for extracting spatial–signal features and a Gates-Controlled Deep Unfolding Network (GCDUN) for iterative refinement. The Addax Optimization Algorithm (AOA) is employed for energy-aware clustering, while Brown Bear Optimization (BBO) fine-tunes hyperparameters to enhance adaptability across varying network conditions. Extensive MATLAB-based simulations in a 100&#xa0;m × 100&#xa0;m network with 100–400 nodes show that LGDU-BBO achieves 99.9% localization accuracy, reduces energy consumption by 25%, and extends network lifetime by 99.8% compared to state-of-the-art methods. Clustering efficiency improves by 15%, and the framework maintains robustness under varying node densities and noisy environments. Comparative analysis against advanced techniques confirms superior performance in accuracy, clustering, energy efficiency, and runtime. The proposed LGDU-BBO framework offers a scalable, efficient, and adaptable solution for large-scale WSN deployments, making it suitable for IoT, smart infrastructure, and industrial automation applications.</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

Advanced localization in wireless sensor networks with attention-guided deep unfolding

  • 期刊论文

  • 发表日期：

    2025年10月1日

查看全文
  • B. Nithya,
  • B. G. Prasanthi

摘要

Accurate localization and energy-efficient clustering are critical for Wireless Sensor Networks (WSNs), especially in dynamic and resource-constrained environments. Traditional methods struggle to balance precision, computational cost, and energy consumption, often leading to reduced scalability and network lifetime. This paper presents a novel Label-aware Gates-Controlled Deep Unfolding Network with Brown Bear Optimization (LGDU-BBO) for high-precision WSN localization. The framework integrates a Label-aware Attention Network (L2AN) for extracting spatial–signal features and a Gates-Controlled Deep Unfolding Network (GCDUN) for iterative refinement. The Addax Optimization Algorithm (AOA) is employed for energy-aware clustering, while Brown Bear Optimization (BBO) fine-tunes hyperparameters to enhance adaptability across varying network conditions. Extensive MATLAB-based simulations in a 100 m × 100 m network with 100–400 nodes show that LGDU-BBO achieves 99.9% localization accuracy, reduces energy consumption by 25%, and extends network lifetime by 99.8% compared to state-of-the-art methods. Clustering efficiency improves by 15%, and the framework maintains robustness under varying node densities and noisy environments. Comparative analysis against advanced techniques confirms superior performance in accuracy, clustering, energy efficiency, and runtime. The proposed LGDU-BBO framework offers a scalable, efficient, and adaptable solution for large-scale WSN deployments, making it suitable for IoT, smart infrastructure, and industrial automation applications.