en.Wedoany.com Reported - Jaiveer Singh, a robotics software engineer leading the NVIDIA Isaac ROS (Robot Operating System) team, is dedicated to building software infrastructure for the physical AI era. His team provides CUDA-accelerated libraries and AI models for developers building autonomous mobile robots, manipulation/grasping systems, and humanoid robots, based on the open-source ROS 2 framework.

Isaac ROS is NVIDIA's open-source software contribution in this field, leveraging the NVIDIA Jetson platform and CUDA libraries to support robot development. The platform supports manipulation/grasping, mobile, and humanoid robots, and provides software packages for perception, object detection, mapping, collision detection, and motion planning. It can run on workstations, the NVIDIA DGX Spark personal AI supercomputer, and NVIDIA Jetson edge systems.

Singh stated that compared to the original Isaac SDK, Isaac ROS is fully modular. Developers can use these packages in combination with existing ROS code, much like assembling LEGO bricks. This modular design enables robot manufacturers to act faster on a foundation that can be inspected, adjusted, and trusted.

Singh believes the primary value of open source lies in giving developers the confidence to build upon it at an early stage and trust its maintainability in the years to come. This confidence is particularly important in the rapidly changing landscape of robotics, especially as humanoid robots have transitioned from science fiction concepts to an active engineering frontier.

Singh's team is working to make Isaac ROS better adapt to this development trend, including providing support for developers using AI agents and for humanoid robot systems requiring end-to-end software stacks. He noted that NVIDIA recognized and began addressing challenges in the robotics field early in the industry, establishing a first-mover advantage.

In Singh's view, open source is a way to share confidence and responsibility. Unlike closed systems, open-source software allows developers to inspect, modify, contribute code, and drive its evolution, turning one company's bug fixes into another's accelerators.