MUMBAI, India, June 30 -- Intellectual Property India has published a patent application (202641075865 A) filed by Vellore Institute Of Technology on June 19, 2026, for A System And Method For Real-Time Multi-Class Intrusion Detection In Iot Networks.
Inventors include S Vinila Jinny; Aryaman Ghaisas; and Syed Zuhaib Ahmed.
The application for the patent was published on June 26, 2026, under issue no. 26/2026.
Abstract: A system and method for real-time multi-class intrusion detection in Internet of Things (IoT) networks is disclosed. The system (100) comprises a network traffic acquisition module configured to obtain network flow data, a feature engineering module configured to perform variance filtering, correlation filtering, and feature ranking, and a normalization module configured to generate normalized feature vectors. A machine-learning inference module generates class probability values corresponding to multiple traffic classes. A threshold optimization module applies independently assigned class-specific threshold values to the generated probabilities and produces threshold-adjusted scores for classification determination. A confidence evaluation module generates a confidence metric and a low-confidence indication when a predefined confidence criterion is not satisfied. The system (100) further comprises application programming interfaces configured for single-flow and batch- flow processing. The disclosed architecture provides a lightweight multi-class intrusion detection framework comprising feature selection, threshold-normalized classification, confidence-aware prediction evaluation, and deployable inference processing for network traffic analysis..
Disclaimer: Curated by HT Syndication.
