MUMBAI, India, Aug. 10 -- Intellectual Property India has published a patent application (202641091306 A) filed by Vels Institute Of Science, Technology & Advanced Studies Vistas; Chennai Tamil; Dr. A. Rajesh; Dr. R. Balakrishna; Mr. N. Udayakumar; Dr. R. Anandan; Mrs. A. Banushri; and Dr. T. Gopalakrishnan on July 28, 2026, for Edge Ai System And Method For Virtual Path Planning And Real-Time Traffic Congestion Management.
Inventors include Dr. A. Rajesh; Dr. R. Balakrishna; Mr. N. Udayakumar; Dr. R. Anandan; Mrs. A. Banushri; and Dr. T. Gopalakrishnan.
The application for the patent was published on July 31, 2026, under issue no. 31/2026.
Abstract: The present invention discloses a smart AI-based system and method for virtual path strategic planning and real-time traffic congestion management, executing entirely on embedded Edge AI hardware at urban intersections. An Electronic Image Stabilization module applies optical-flow-based affine compensation to reduce camera jitter, and an Adaptive Frame Rate Control module switches between high and variable frame-rate processing based on a Motion Intensity Index, reducing computational load during low-traffic periods. A lightweight CNN engine detects and classifies cars, motorcycles, auto-rickshaws, buses/trucks, and pedestrians in real time. The core contribution is a Virtual Path Strategic Planner representing the intersection as a congestion-weighted Dynamic Virtual Grid over which parallelized multi-agent A* and Dijkstra path planning executes concurrently, employing a cell reservation protocol for collision avoidance and sub-second replanning. Congestion indices derived therefrom drive dynamic traffic signal phase adjustment without dependence on cloud connectivity.
Disclaimer: Curated by HT Syndication.
