A configuration combining the Turtlebot3 with a Jetson Nano for pose estimation and gesture control
There are several folders at the base of the trunk. One for each of the computer systems involved in the project.
- Jetson Nano Handles the computer vision and machine learning aspect of the project.
- Raspbian Handles the interpreting of the poses.
- OpenCR Handles the ROS (Robot Operating System) tasks.
- Ubuntu16.04 Acts as the x86_64 host computer.
The Parser subfolder/project in the Raspbian folder is for parsing .json files that are output by OpenPose, written in C#/Mono.
Activating the system works as follows:
- Turn on the Jetson Nano, run the ./start_script (this also assigns it an IPv6 Address)
- Turn on the Turtlebot3 (the switch on the OpenCR board)
- Either SSH into or directly access the Raspberry Pi and start the parser application ("mono parser.exe")
Now, being in front of the camera with one's right arm stretched outward will make the Turtlebot3 drive forward. Be careful as to not have it drive off of anything.