This work is part of a larger project, it is designed to be used with the Authoring Interface.
This guide is designed for system using a Panda robot controlled through a graphical user interface, possibly remotely. The physical system requires a panda outfitted with an azure kinect on the end-effector using the printed mount provided in the meshes folder and ATI Axia80-M20 6-axis force torque sensor. The physical system requires a computer running Ubuntu 18.04 for the Kinect Ros driver, a computer running Ubuntu 20.04 for the Object Localisation. The present code can be run on either platform. The panda low-level controller and interface with the ATI can be found here. More information to recreate the workspace can be requested by contacting esenft@wisc.edu.
However, we also provide a simulated system that can be run on a single Ubuntu computer without the robot.*
If not yet installed, start by installing ROS (tested with ROS Noetic, but other versions might work as well).
Dependencies:
Recommend to use catkin_tools to compile. Then build with:
> catkin build authoring
The correct ROS_IP environment variable needs to be set before use.
First, start roscore in a dedicated terminal.
roslaunch authoring all.launch only_virtual:=true
This system is designed to be interfaced with the Task-Level Authoring which can be run using QtCreator or served on a website.
To start a condition, run the related code through QtCreator. Of note, the default setting for this release is simulation, but when using the real system (roslaunch authoring all.launch), the line property bool simu: true in each main.qml file needs to be set to false.