Sample applications for ROS combined with a (simulated) Franka Emika Panda robots

⚠️ This repository must be checked out in a catkin workspace containing the franka_description, panda_moveit_config, and panda_simulation ROS packages provided here. For more information and installation instructions, please consult the main project.

Contains different examples for the planning and execution of robotic motions with MoveIt in ROS.

Example Applications

Please note that all examples must be run in a built catkin workspace from within a shell with all ROS paths properly set up (i.e., after running source devel/setup.sh).

Minimal "Hello World" Example

This example of an execution of a simple motion.

• Command: roslaunch sample_applications sample_minimal_simulation.launch
• Expected results:
  • A Gazebo window opens in which the executed robot motion can be observed.
  • The motion starts immediately after the the system has loaded.

Execution of a Simple Motion with Tooling

This is the same example as the first one, but this time, additional tooling is loaded.

• Command: roslaunch sample_applications sample_simple_simulation.launch
• Expected results:
  • A Gazebo window to observe the executed robot motion opens
  • An rqt window to show log messages opens
  • An RViz window to see the plan and trigger the motion open.
  • The motion does not start automatically, instead it must be triggered by pressing next in the RvizVisualToolsGUI within RViz

Execution of a Simple Motion Constrained by a Blocking Object

In this example, an object can be observed in the RViz planning window. This object constrains the motion of the robot, such that the robot moves around it.

• Command: roslaunch sample_applications sample_constraint_simulation.launch
• Expected results:
  • A Gazebo window to observe the executed robot motion opens
  • An rqt window to show log messages opens
  • An RViz window to see the plan and trigger the motion open.
  • The motion does not start automatically, instead it must be triggered by pressing next in the RvizVisualToolsGUI within RViz

Execution of a Velocity Constraint Cartesian Trajectory

• Command: roslaunch sample_applications timed_cartesian_simulation.launch
• Expected results:
  • A Gazebo window to observe the executed robot motion opens
  • An rqt window to show log messages opens
  • An RViz window to see the plan and trigger the motion open.
  • The motion does not start automatically, instead it must be triggered by pressing next in the RvizVisualToolsGUI within RViz

Execution of a Pick-and-Place Task

• Command: roslaunch sample_applications pick_place_simulation.launch
• Expected results:
  • A Gazebo window to observe the executed robot motion opens
  • Two tables an a "stick" on top of one of them are spawned in Gazebo
  • The Robot positions itself behind the stick and grasps it from the side.
  • It then moves to the other table and places it on top of it.
• Remarks and potential issues:
  • The grasping in not done using the physical properties of the object (which would require extensive tweaking of the surface properties of it), but using a Gazebo plugin grasp fix.
  • The process of this tutorial is a bit fragile. It may happen that the gripper fails to grab the object, that it slips out of the gripper during the motion or that is is not placed correctly on the second table (and thus falls down). We believe that these problems might be reduced by tweaking of some properties, but in the end, all of the problems can also happen during an actual execution, so we have to live with them.
  • If a problem like this occurs, be thankful that it was just a simulation and simply restart.

Planning and Simulation based on RViz

This is an example to use RViz and Gazebo interactively. Motions can be planned and executed from the RViz GUI.

• Command: roslaunch sample_applications simulation.launch
• Expected results:
  • RViz window to control the robot
  • Gazebo window to observe the executed robot motion
  • rqt to show log messages
  • The robot does not move automatically, but can be controlled with RViz (cf. the tutorial)