This is the initial release of Intel® Robot Development Kit open source project, which enables developers to easily and efficiently create, customize, optimize, and deploy a robot software stack to an Autonomous Mobile Robot (AMR) platform based on the Robot Operating System 2 (ROS2) framework. The Robot DevKit supports Ubuntu 18.04 and has been verified on Intel NUC (CPU: Intel i7-6700HQ @2.60GHz, Mem:16G) and ADLink Neuron Board (CPU: Intel i7-6700TE @2.40GHz, Mem:16G) with TurtleBot3 Waffle chassis. Installation includes the software components below that are made available in your development environment to build your own Robot.
-
ROS2 core for key packages of ROS2 framework
-
ROS2 Intel® OpenVINO™ with CPU, GPU and VPU acceleration for open visual inference and neural network optimization providing easy to use face detection, person identification, age and gender recognition, and more!
-
ROS2 Object Analytics with Intel® Movidius NCS™ for object detection, tracking and 3D localization.
-
ROS2 navigation and SLAM for map generation and robot navigation with the map
-
Rviz2 visualization tool and gazebo-ros2 simulator
-
Intel® RealSense™ for 3D vision, point cloud creation, SLAM and navigation
-
Tutorials for key components
The Robot DevKit provides off-the-shelf features to developers, like avoiding obstacles while navigating, using models in OpenVINO model zoo for object or people face detection, optimizing your models to get improved neural network performance on a variety of Intel® processors (CPU, Intel® Processor Graphics, VPU), etc. In the near future, we’ll be adding robot arm features with motion planning and control, perception for grasping, hand-eye calibration, etc. Stay tuned for more features!
Your feedback and suggestions are highly appreciated. Feel free to provide feedback or report issues in github.