MUMBAI, India, Sept. 14 -- Intellectual Property India has published a patent application (202641106137 A) filed by Sri Eshwar College Of Engineering on September 03, 2026, for Asthra Xd - Monocular Depth Estimation And Obstacle Detection System For Autonomous Robots.
Inventors include Dr. S. Saranya; Ms. V. Dhivya; and Ms. T. Dhanushya.
The application for the patent was published on September 11, 2026, under issue no. 37/2026.
Abstract: ASTHRA XD is a real-time robotic perception and navigation system that integrates path tracking, object detection, and depth estimation to enhance autonomous operation and situational awareness. The system uses a single Web camera and deep learning models, such as MiDaS v3.1, running on an NVIDIA Jetson Orin Nano platform, to produce precise depth maps. These maps are then combined with YOLOv11-based detections to detect obstacles and determine their distances in real time. A DeepSORT framework combined with Kalman filtering enables efficient path prediction and motion analysis, guaranteeing accurate identity assignment and smooth trajectory tracking of dynamic objects. The system includes a voice alert module powered by VOSK to deliver immediate spoken warnings like "Required object detected on the table," for safety feedback. In addition to facilitating modular deployment and effective communication between perception and navigation components, the system's ROS 2-based architecture also supports cooperative robotic operation by enabling multiple robots to share obstacle and path information via wireless interfaces like Wi-Fi to enable coordinated navigation and intelligent decision making. ASTHRA XD has been created as a low-cost substitute for LiDAR-based perception systems. Its affordability, scalability, and real-time responsiveness make it appropriate for multi-robot environments, autonomous robots, service robotics, and warehouse automation.
Disclaimer: Curated by HT Syndication.
