IIIT-H, Ati Robotics Test Vision-Based Robot Navigation

IIIT-Hyderabad and Bengaluru-based Ati Robotics have demonstrated an AI-powered navigation system. It enables autonomous mobile robots to navigate factory floors, using camera images instead of LiDAR.

The demonstration is part of a three-year research collaboration between the two organisations. The robot identified and navigated towards a target object inside Ati Robotics’ factory using visual information and an AI-generated model of its surroundings.

The technology uses photographs obtained during a previous site inspection. The scientists use a 3D artificial intelligence model to compare two photos in order to figure out how the factory is structured and provide the robot with directions to its destination. The method is based on the combination of vision-language technologies with image navigation models.

Unlike LiDAR devices, which measure distances using laser rays. They create a map of the environment. But this technology relies on visual information. It is used by robots to comprehend the environment. In addition, the technique includes elements of human navigation, such as humans use memory, visual information, and movements while navigating in space.

Most systems for internal mobility use  LiDAR-based navigation for industrial robots and this research provides a different approach to movement once the performance holds across  different factory layouts. 

Image source: IIIT -H