MUMBAI, India, June 26 -- Intellectual Property India has published a patent application (202641072289 A) filed by Vellore Institute Of Technology on June 10, 2026, for Adaptive Multi-Modal Behavioural Threat Assessment Apparatus For Residential Entry Points Using Physics-Grounded Real-Time Intent Scoring And Constrained Personalized Learning.
Inventors include Archana Tamizharasan; Kr Balasubramanian; Naziya; and Gouri Shiva.
The application for the patent was published on June 19, 2026, under issue no. 25/2026.
Abstract: The present invention relates to residential security and behavioural threat assessment. The invention discloses an adaptive multi-modal behavioural threat assessment apparatus configured for real-time monitoring of residential entry points. The apparatus integrates behavioural landmark extraction, aggression probability estimation, hazardous object identification, authorized identity attenuation, and hardware interaction monitoring into a unified behavioural fusion framework. Behavioural indicators including loitering persistence, directional oscillation, stop-go movement, and doorbell interaction persistence are transformed into normalized signals and fused to generate a continuous behavioural intent score. Deterministic override logic elevates threat classification upon occurrence of safety-critical conditions. The invention further incorporates adaptive behavioural personalization using homeowner-labelled events and contextual behavioural weighting for different temporal conditions. The apparatus enables automated threat classification, event recording, behavioural summarization, and residential response activation with low-latency edge-based deployment.
Disclaimer: Curated by HT Syndication.
