MUMBAI, India, Feb. 6 -- Intellectual Property India has published a patent application (202541123820 A) filed by Nandha Engineering College, Erode, Tamil Nadu, on Dec. 9, 2025, for 'an intelligent surveillance rover integrating real time object detection with llm based semantic analysis.'
Inventor(s) include Rajasekaran T; Udhay Karthik C; Pravin Kumar S; and Sachin K.
The application for the patent was published on Feb. 6, under issue no. 06/2026.
According to the abstract released by the Intellectual Property India: "Modem security is often hindered by old technology: fixed cameras that can only stare at one spot and require constant human attention to spot trouble. These stationary systems create blind spots and, crucially, can't tell the difference between a person harmlessly walking by and a genuine threat, lacking any real context or understanding of a situation. Our project tackles this by introducing ROYE, a compact, mobile surveillance robot ROV-E is built on an affordable Raspberry Pi 5 platform using the flexible ROS 2 framework, and its primary goal is to bring intelligence and mobility to security monitoring. ROY -E's core innovation is its powerful combination of vision and language. It uses the cuttingedge YOLOv8 model for high-speed, on-device object detection, quickly identifying items like people, bags, and obstacles in real time. Instead of just reporting coordinates, ROV-E sends this detection data to Google's Gemini API, a cloud-based Large Language Model, which translates the raw sensor data into rich, natural language descriptions. This crucial step allows ROY -E to move beyond simple motion alarms to provide genuine scene understanding and detect uncharacteristic activity, giving operators instant, meaningful context about what the rover is seeing. For movement, ROV-E is designed for tight indoor spaces, utilizing an omnidirectional Mecanum wheel base that allows it to move forward, backward, sideways, and diagonally with high agility. Users control the rover remotely through a simple, responsive React-based web interface that also displays the AI's real-time semantic alerts. This architecture delivers dynamic coverage and intelligent semantic feedback as a low-cost, adaptable alternative to expensive, centralized security systems. We believe. this lays the groundwork for the next stage of robotic security, with plans to integrate LiDAR-based SLAM to transform ROV-E into a fully autonomous patrol agent."
Disclaimer: Curated by HT Syndication.
