Graph Neural Network-Driven Networking for Robust Industrial Wireless Sensor Networks

作者：Yuhan Su, Yu‐Chen Lin, Xinqin Liao, Zhong Chen, Tingzhu Wu · 发表于：IEEE Transactions on Industrial Informatics · 年份：2025 · DOI：10.1109/tii.2025.3632658 · 被引用次数：4 · 研究领域：Energy Efficient Wireless Sensor Networks、IoT and Edge/Fog Computing、Network Time Synchronization Technologies

Industrial wireless sensor networks (IWSNs) play a critical role in enabling real-time monitoring and intelligent automation in modern industrial applications. However, maintaining reliable communication and efficient data transmission in dynamic and interference-prone environments remains a significant challenge. To address these limitations, this article proposes a graph neural network (GNN)-driven networking approach for IWSNs, designed to enhance communication robustness and optimize data processing. Our approach incorporates a minimum capacity constraint and a trainable slack parameter, enabling adaptive network configuration in response to changing conditions. By modeling the network topology as a graph, we formulate a device-centric joint node selection and power allocation (JNP) strategy, leveraging GNNs for real-time decision-making. Simulations benchmark the proposed method against state-of-the-art methods, showing up to a 70% average increase in fifth percentile rate across various network conditions. These results highlight the effectiveness of the proposed JNP strategy in improving IWSN performance for industrial applications.