Key Advantages of this platform are:
Pre-Populated Robots and Sensors
ROS/ROS 2.0 Support
Custom ROS messages and URDF/MJCF are now open sourced
Scalable Synthetic Data Generation
SimReady Assets
NVIDIA Isaac ROS
We use DNNetworks a foundation model to delivering high-performance perception and NVIDIA-accelerated computing hardware acceleration to ROS-based robotics applications
cuMotion for Robot Manipulation
We Integrate NVIDIA® CUDA®-accelerated path planning for robotic arm-related tasks such as collision detection using robot masking and integrated trajectory optimization using the MoveIt motion-planning framework
FoundationPose
We are working with state-of-the-art foundational model for 6D pose estimation to detect novel objects
NVIDIA Isaac Transport for ROS (NITROS)
nvBlox uses RGB-D data to create a dense 3D representation of the robot's environment, this helps us in narrowing down the error margins as it generates a temporal costmap for the navigation stack
Stereo Perception
We use Stereo Perception, that are DNN-based GEMs designed to help roboticists with common perception tasks
We use Enhanced Semi-Supervised (ESS) stereo disparity which is a DNN for stereo camera disparity prediction and Bi3D is a DNN for vision-based proximity detection
Comments