Ditching LiDAR: IIIT-H, Ati Robotics Demonstrate AI Vision-Based Robot Navigation
In the partnership, which began in 2023, researchers explored how robots could navigate using visual cues and spatial relationships, much as people recognise familiar landmarks to find their way.
HYDERABAD: An autonomous mobile robot has navigated to a target on a factory floor using camera images and AI-based 3D reconstruction, without relying on the conventional LiDAR maps commonly used in robotic navigation. The demonstration resulted from a practical test of research developed by the Robotics Research Center at the International Institute of Information Technology Hyderabad (IIIT-H), in collaboration with Ati Robotics.
In the partnership, which began in 2023, researchers explored how robots could navigate using visual cues and spatial relationships, much as people recognise familiar landmarks to find their way.
“LiDAR systems map surroundings and help robots estimate their position through simultaneous localisation and mapping (SLAM). The IIIT-H team instead investigated whether robots could use images, recognisable landmarks and artificial intelligence to locate themselves and plan routes without building dense, detailed maps,” said Prof. Madhava Krishna of the Robotics Research Center.
He explained that the approach represents environments through a smaller set of landmarks. In a warehouse, these could be locations near a door or opposite an elevator, rather than a map containing detailed measurements of every part of the space.
The researchers also tested image-based navigation, in which an operator identifies a destination by pointing to a pixel in an image rather than entering precise coordinates. The system compares images to identify corresponding points in the real world, estimate their spatial relationships and assess possible routes.
Prof. Sourav Garg of the centre said the research had resulted in papers presented at international forums, including the IEEE International Conference on Robotics and Automation, NeurIPS and the Conference on Robot Learning.
IIIT-H has transferred code to Ati Robotics for testing and deployment in warehouse robots. The company has adapted parts of the research into OccupancyReact, a LiDAR-based variant closer to its existing systems.
Ati Robotics said its current focus was improving the final stages of robot movement, including docking, approaching pallets and manoeuvring through narrow spaces, where even small-time savings can accumulate across a working shift.