MUMBAI, India, July 24 -- Intellectual Property India has published a patent application (202641085326 A) filed by Velammal Engineering College on July 12, 2026, for A Federated Learning Framework For Privacy-Preserving Anomaly Detection In Distributed Internet Of Things (iot) Networks.
Inventors include Dr. P. Visu; Mrs. A. Prema; Mrs. P. V. Raja Suganya; and Mrs. R. Kavitha.
The application for the patent was published on July 17, 2026, under issue no. 29/2026.
Abstract: The present system discloses a Federated Learning Framework for Privacy-Preserving Anomaly Detection in Distributed Internet of Things (IoT) Networks. The framework enables decentralized collaborative learning by allowing IoT devices and edge nodes to train anomaly detection models locally while retaining raw data on-device. Encrypted and compressed model updates are transmitted to a secure aggregation server employing adaptive federated weighted aggregation based on model accuracy, data quality, device reliability, and trust scores. The invention further incorporates differential privacy, secure aggregation, trust-based client selection, gradient compression, and explainable artificial intelligence to enhance privacy, communication efficiency, robustness, and interpretability. The proposed framework provides accurate, scalable, and low-latency anomaly detection suitable for healthcare, industrial IoT, smart cities, transportation, and other distributed IoT applications while significantly reducing privacy risks and communication overhead compared with conventional centralized approaches.
Disclaimer: Curated by HT Syndication.
