Intellectual Property India Publishes Patent Application for 'Machine Learning Driven Intrusion Detection System For Secure Internet Of Vehicles' Filed by Agni College Of Technology

MUMBAI, India, May 29 -- Intellectual Property India has published a patent application (202641050498 A) filed by Agni College Of Technology, Chennai, Tamil Nadu, on April 21, for 'machine learning driven intrusion detection system for secure internet of vehicles.'

Inventor(s) include Kannan R; Krishna SB; Akilan A; Kartheeswaran V; and Balaji S.

The application for the patent was published on May 29, under issue no. 22/2026.

According to the abstract released by the Intellectual Property India: "This invention presents a Machine Learning Driven Intrusion Detection System (ML-IDS) for the Internet of Vehicles (IoV), providing real-time detection, classification, and response to cybersecurity threats across V2V, V2I, and V2X vehicular communication networks. The system integrates a multi-stage preprocessing pipeline for CAN bus, DSRC/C-V2X, and OBD-II data streams with an ensemble classification engine combining Random Forest and XGBoost for known attack detection, and an LSTM-Autoencoder for zero-day anomaly identification via adaptive reconstruction error thresholding. Federated Learning coordination layer enables privacy-preserving distributed model training across vehicle fleets using FedAvg aggregation with Homomorphic Encryption and Differential Privacy guarantees, eliminating centralized data exposure. An Explainable AI subsystem generates SHAP-based feature attribution maps and human-readable intrusion alerts compliant with ISO 21434 automotive cybersecurity standards. The proposed ML-IDS achieves 99.1% detection accuracy across the OTIDS, VeReMi, and CICIDS-2017-V2X benchmark datasets, outperforming Decision Tree, SVM, CNN-IDS, and GRU-Anomaly baseline models across precision, recall, F1-score, and AUC-ROC. The system detects a comprehensive IoV threat taxonomy including DDoS flooding, replay attacks, Sybil attacks, MitM intrusions, CAN bus injection, GPS spoofing, and false data injection with sub-10ms edge inference latency. Deployed as an embedded IDS agent within AUTOSAR Adaptive Platform ECUs and RSU infrastructure, the system delivers scalable, privacy-compliant, and safety-critical intrusion detection for next-generation connected and autonomous vehicle ecosystems."

Disclaimer: Curated by HT Syndication.