Cogniteam, a robotics company and creator of the Nimbus low-code robot development operating system, has integrated the latest NVIDIA® Jetson AGX Orin™ platform into its existing robot software packets. As a Preferred member of the NVIDIA Partner Network, Cogniteam had early access to the AGX Orin developer kit, which allowed the company to prepare a streamlined process for the next generation of edge robot capabilities.
Creating greater ease of development, testing, and deployment has cut the development times for robotics software dramatically. Yet, companies still face challenges when integrating new technologies. This is especially true for those who are looking to deploy a brand new system-on-module, such as the recently released Jetson AGX Orin.
To address the concerns of the industry and ease deployment, Cogniteam was able to include all the necessary Jetson Orin drivers and software components into its Nimbus software platform. Using Nimbus, it is now much quicker to get a robot with Jetson Orin up and running for inspection, maintenance, sanitization, service, and various data-intensive scenarios that require a high degree of autonomy. With a few clicks from the browser, developers are now able to drag and drop ready-made NVIDIA Isaac software development kit GEMs as Nimbus dockerized components. Users also have better support for ROS2 components and can upgrade them remotely, from anywhere in the world. “The NVIDIA Jetson AGX Orin is a huge leap for manufacturers and developers who want greater AI capabilities on their edge robots,” said Dr. Yehuda Elmaliah, Co-Founder & CEO of Cogniteam. “We were pleased to see that the development of this new module integrates seamlessly with the Nimbus Operating System, enabling deployed robots to handle greater AI computational abilities.”
As more robots gain cloud connectivity, AGX Orin’s greater computing power will meet the market’s demand for outfitting robots with greater autonomous capabilities, including AI decision making. “Relying on Nimbus with Jetson AGX Orin, we dragged, dropped, and ran multiple AI models simultaneously on the robot,” said Ari Yakir, VP of Research & Development at Cogniteam. “These skills included localization, gesture recognition, skeleton recognition, object recognition, simultaneous localization and mapping (SLAM), and beyond.”