MUMBAI, India, June 26 -- Intellectual Property India has published a patent application (202641069394 A) filed by Psg Institute Of Technology on June 03, 2026, for A Multi Modal Framework For Driver Behavior Monitoring And Assessment Using Obd Ii Vehicle Data For Fleet Management.
Inventors include Dr. J. Niresh; Dr. S. Neelakrishnan; Mr. Ohm Pranav Dv; Azhagan Vivekanandhan Nr; and Anirudh K.
The application for the patent was published on June 19, 2026, under issue no. 25/2026.
Abstract: ABSTRACT OF THE INVENTION "A MULTI-MODAL FRAMEWORK FOR DRIVER BEHAVIOR MONITORING AND ASSESSMENT USING OBD-11 VEHICLE DATA FOR FLEET MANAGEMENT" The increasing demand for intelligent, accessible, and scalable data-driven applications has driven the adoption of cloud-native machine learning and large language model (LLM) architectures. In this work, we present a cloud-hosted analytical system that integrates supervised machine learning, cloud databases, and LLM-based reasoning to deliver real-time insights through a web-based interface. A trained classification model generates predicted classes from incoming data, which are securely stored in a cloud-based PostgreSQL database using Supabase. A Streamlit-powered user interface dynamically retrieves the required class from the cloud database and assigns it as a system-side variable for prompt engineering. This structured input is then processed by an LLM to generate detailed analytical reports and data visualizations tailored to the predicted class. The application is deployed as a Streamlit-based cloud web application, accessible through standard web browsers on both desktop and mobile devices. By leveraging cloud infrastructure, the system ensures real-time availability, scalability, and seamless access from any location with an internet connection. Experimental deployment demonstrates reliable end-to-end data flow, low-latency retrieval, and effective generation of contextual analytics. The proposed architecture highlights a practical and extensible framework for integrating ML inference, cloud data management, and LLM-driven analysis in modern intelligent applications. Figure 2 Shows tstoring t
Disclaimer: Curated by HT Syndication.
