MUMBAI, India, Aug. 31 -- Intellectual Property India has published a patent application (202641099015 A) filed by Koneru Lakshmaiah Education Foundation on August 14, 2026, for Deeplift-Ae Driven Granular Feature Extraction For Intelligent Detection Of Emerging Vehicular Network Threats.
Inventors include Dhanda Supriya; and Anne Venkata Praveen Krishna.
The application for the patent was published on August 28, 2026, under issue no. 35/2026.
Abstract: ABSTRACT: Title: DeepLIFT-AE Driven Granular Feature Extraction for Intelligent Detection of Emerging Vehicular Network Threats The present invention relates to an Explainable Artificial Intelligence based Intrusion Detection System (XAI-IDS) for detecting emerging threats in vehicular communication networks. The system receives and preprocesses vehicular traffic, including Controller Area Network (CAN) data, and employs an integrated Fuzzy Probability Clustering with SHAP (FPC-SHAP) module for uncertainty-aware feature extraction and interpretable feature contribution analysis. The extracted features are supplied to a DeepLIFT-enhanced Autoencoder (DeepLIFT-AE) configured to generate latent representations, determine feature relevance and rank significant features while reducing redundant information. The ranked features are subsequently processed by a Local Interpretable Model-agnostic Explanations based Globally Randomized Ensemble (LIME-GRE) classifier configured to classify traffic into benign, Denial-of-Service (DoS) and spoofing classes. LIME generates local surrogate explanations using perturbed neighbourhood samples, while aggregated feature contributions provide global interpretability. The integrated architecture improves detection robustness, feature transparency, classification traceability and computational efficiency for real-time IoV cybersecurity applications.
Disclaimer: Curated by HT Syndication.
