Author: Tobit Flatscher (March 2023)
Sometimes you come across some interesting old work that you would like to replicate. Sometimes a publication comes with open-source code but likely your ROS version will not necessarily match the one that was used in the publication. For me that was the case when reading the publication about Affordance Templates published by Traclabs (you might know them from trac-ik) at ICRA 2015 (see image below). And even if you had an old computer with ROS Indigo (released in 2014!) laying around you would not want to trash that computer with all the dependencies just for trying this one workspace out. This is again where Docker as a technology comes in really handy.
The following example will shows how Docker can be useful for getting legacy code up and running on your machine. This example includes the packages from the Traclabs Bitbucket with the commits specified in this workspace. Potentially one could also get this workspace up and running in ROS Fuerte by following this guide and using the deprecated packages from the NASA Bitbucket (nasa_common, nasa_robodyn, nasa_r2_simulator and nasa_r2_common).
Please make sure that you cloned this repository with the --recurse-submodules option as follows
$ git clone --recurse-submodules https://github.com/2b-t/docker-for-robotics.gitelse the modules from Traclabs will be empty. In case you did not you will have to run
$ git submodule update --initOpen a terminal on your host machine and allow the user root to stream onto your display. This way the Docker (which is using root as a user) can open graphic user interface on your host machine:
$ xhost +local:rootOpen your Docker either manually or by using Visual Studio code as described in the main guide. Once inside the Docker perform the following steps:
$ source /opt/ros/indigo/setup.bash
$ catkin build
$ source devel/setup.bash
$ roslaunch at_r2_bringup demo.launchThis should open two new windows, one for the Gazebo simulation, the other for the Rviz visualization (see screenshot above). The Rviz visualization should contain two panels: The Moveit motion planning panel to the left and the affordance templates panel to the right. Crucial for the latter is the interactive marker that is displayed. Furthermore any object inserted into the scene should be visible from the Asus camera plug-in.
Inside Moveit you can only plan for the head. To do so select a desired goal state from the section Query in the Planning tab. Then click Update and finally plan and execute the motion with Plan and Execute.
The affordance templates panel on the right can be used for planning arm motions. As a first step click Refresh All in the Server tab. This should bring up all the available motion templates in the Library tab below. For inserting a new template double click on its icon. This will insert a virtual marker into the scene (if it does not appear make sure that you did not deactivate the interactive marker in the Displays panel!). You can adjust the parametrized motion by moving and rotating the interactive marker. This is crucial as the Robonaut will not be able to plan and execute a motion if it is out of reach. Finally in the Controls menu you can plan and execute the parametrized motion once inserted: The arrow buttons on the bottom allow you to plan the templates step by step. Click Execute Plan to then execute the planned motion. When activating the checkbox Execute on Plan stepping the motion will also result in the execution of the plan.