Isaac is NVIDIA’s open platform for intelligent robots. The Isaac SDK provides a large collection of powerful GPU-accelerated algorithm GEMs for navigation and manipulation. Isaac SDK Engine is a framework to easily write modular applications and deploy them on a real robot. Isaac SDK comes with various example applications from basic samples that show specific features to applications that facilitate complicated robotics use cases. Isaac SDK also works hand-in-hand with Isaac SIM, which allows for development, testing, and training of robots in a virtual environment.
Robotics combines many different disciplines, including low-level hardware drivers, safe planning algorithms, fast and accurate computer vision, deep neural networks, and high-level artificial intelligence. Suceeding in these disciplines often requires years of expertise.
Isaac SDK comes with a collection of high-performance algorithms, also called GEMs, to accelerate the development of challenging robotics applications. For example, Isaac provides planning and perception GEMs for navigation and manipulation use cases. GEMs also provide support for key hardware components and robotic peripherals.
Isaac SDK provides various sample applications, which highlight features of Isaac SDK Engine or focus on the functionality of a particular Isaac SDK GEM. These sample applications are good starting points for learning Isaac.
The Isaac SDK is meant for development of applications for complicated use cases like a delivery robot. The Carter application gives you a starting point for building your own delivery robot. Carter can drive to a goal location, patrol a building, or bring you popcorn. The Carter navigation stack is based on a Lidar.
To start on a smaller scale, see the Kaya application. Isaac includes instructions on how to build your own robot and bring it to life with artificial intelligence.
Isaac SDK is also supported by a rich ecosystem, and Isaac SDK Engine connects Isaac GEMs to existing packages like OpenCV, ROS, PCL, and others.
Isaac SDK includes the Isaac SDK Engine, a feature-rich framework for building modular robotics applications. With Isaac, you can build an application out of small components, which pass messages between each other and can be configured to your custom use case.
Isaac SDK comes with toolchains based on the Bazel build system for building and deploying
applications. You can build and run applications with a command as simple as
bazel run //apps/samples/realsense_camera. All external dependencies are pulled
automatically to your system without any additional setup. The Setup section of this
document explains the few steps necessary for getting started.
Isaac SDK Engine fully supports NVIDIA GPUs and CUDA, TensorRT, NPP and other frameworks that allow you to build the fastest robotics application. It is optimized for NVIDIA hardware like Jetson Xavier, Jetson TX/2, Jetson Nano, and workstation GPUs. You can deploy a cross-compiled application optimized for your platform with a single, short command line script.
Isaac Sim applies the NVIDIA Omniverse™ platform to robotics with state-of-the-art RTX graphics and GPU-accelerated PhysX simulation. You can use Omniverse Connect to export engineering models from popular tools into the Omniverse Nucleus server, and build your simulation experiments using the Isaac Sim editor. You can learn more about Isaac Sim on the Omniverse documentation page.
Isaac SDK also provides support for Unity3D as a simulation backend for navigation and perception training, with support for the Unity3D high-definition render pipeline (HDRP).