MUMBAI, India, April 17 -- Intellectual Property India has published a patent application (202641043255 A) filed by Vignan'S Foundation For Science, Technology & Research, Guntur, Andhra Pradesh, on April 4, for 'secure federated learning framework using trust evaluation for internal attack detection in vehicular networks.'
Inventor(s) include Dr. DS Bhupal Naik; and Dr. Venkatesulu Dondeti.
The application for the patent was published on April 17, under issue no. 16/2026.
According to the abstract released by the Intellectual Property India: "A trust-based secure federated learning (FL) system has been proposed for intelligent vehicular networks to detect and mitigate internal attacks as shown in figure-1. Each client vehicle collects and preprocesses vehicle-to-vehicle (V2V) communication data, trains a local machine learning model, and securely transmits encrypted model updates to a central server using a public key infrastructure (PKI). The central server decrypts updates, detects anomalies via density-based clustering, and computes trust scores based on direct, historical, and indirect trust. Updates from vehicles exceeding a predefined trust threshold are aggregated to generate a global model, which is securely redistributed for iterative retraining. Dynamic trust evaluation incorporates neighbouring opinions and confidence-based weighting to reduce malicious influence. Experimental validation using VeReMi, CICIDS2017 and AWID datasets demonstrates superior detection accuracy, precision, and resilience under multiple attack scenarios, while maintaining packet reach ratio (PRR) and network throughput. The framework effectively ensures secure, reliable, and trustworthy model updates in dynamic vehicular environments."
Disclaimer: Curated by HT Syndication.
