Heterogeneous Data Fusion and Anomaly Detection in Industrial IoT Systems Using Spatio-Temporal Graph Neural Networks

This paper proposes a novel framework for heterogeneous data fusion and anomaly detection in Internet of Things (IoT) systems using spatio-temporal graph neural networks (GNNs). The framework addresses the challenges of integrating multi-source data and capturing complex spatiotemporal dependencies in industrial environments. A multi-layer GNN architecture with a spatio-temporal attention mechanism is developed to learn hierarchical representations of industrial entities and their interactions. Extensive experiments on real-world IoT datasets demonstrate the superiority of the proposed approach over state-of-the-art methods in terms of accuracy, precision, recall, F1-score, and AUC-ROC. The framework shows robustness to parameter changes and scalability to large-scale datasets, making it suitable for real-time anomaly detection in industrial settings. This work contributes to the advancement of intelligent and reliable industrial operations in the context of Industry 4.0.