This is a set of utilities and examples for control Panda Franka Emika using Moveit.

• Getting started
  • Dependencies and building
• Files
  • data_manager
  • panda
    • panda_arm
    • panda_gripper
    • panda
  • exceptions
  • colors
  • python
• Data
• Examples
  • console
  • pick_and_place
  • teleoperation
    • teleop_talker
    • teleop_listener
• Visual Studio Code
• Author
• License

This package was tested with real robot with:

• Ubuntu 20.04 LTS Focal Fossa
• ROS noetic
• libfranka 0.8.0
• franka_ros 0.7.1
• panda_moveit_config 0.7.5

Install ROS

Install Catkin

Remove possible conflicts:

sudo apt remove ros-noetic-*franka*

Install c++ tools:

sudo apt install clang-format cloc doxygen

Prepare Workspace:

cd

mkdir panda_ws

cd panda_ws/

mkdir lib src

Get libfranka:

cd ~/panda_ws/lib/

sudo apt install build-essential cmake git libpoco-dev libeigen3-dev

git clone --recursive https://github.com/frankaemika/libfranka

cd libfranka/

git checkout 0.8.0

git submodule update

mkdir build

cd build

cmake -DCMAKE_BUILD_TYPE=Release ..

cmake --build .

Get franka_ros:

cd ~/panda_ws/src/

git clone --recursive https://github.com/frankaemika/franka_ros

git checkout 0.7.1

cd ..

rosdep install --from-paths src --ignore-src --rosdistro noetic -y --skip-keys libfranka

Get panda_moveit_config:

cd ~/panda_ws/src/

sudo apt install ros-noetic-moveit

git clone https://github.com/ros-planning/panda_moveit_config.git -b melodic-devel

Get panda_controller:

cd ~/panda_ws/src/

pip3 install --user pyyaml

git clone https://github.com/Envq/panda_controller.git

Building:

cd ~/panda_ws/

catkin config -j$(nproc) --extend /opt/ros/${ROS_DISTRO} --cmake-args -DCMAKE_BUILD_TYPE=Release -DFranka_DIR:PATH=$HOME/panda_ws/lib/libfranka/build

catkin build

echo 'source ~/panda_ws/devel/setup.bash' >> ~/.bashrc

source ~/.bashrc

Use this command for genere the doxygen documentation:

roscd panda_controller

doxygen doxygen.cfg

This file contains the functions necessary to save and read the data in the appropriate yaml files.

This file is a wrapper for easy use of moveit to control the Panda robot, designed for pick and place tasks.

• panda_arm

  In this file are defined all the method for control the arm.

• panda_gripper

  In this file are defined all methods for controlling the gripper with client action for franka_gripper. If you set REAL_ROBOT=False Then you will be able to see in simulation (without having the robot connected) the movement of the gripper.

• panda

  This file contains panda_arm and panda_gripper providing getters to use them. It also adds the ability to create and reset scenes and pick and place. Note: the collision area of the robot hand is too large, so it is not possible to see in RVIZ the pick and place with the stl of the object.

A try catch system has been implemented to get more information about possible errors and exceptions are defined in this file.

This file contains the definitions to modify the colors and format of the stdout. Wrap your output string with them.

The same files are also available for python3 with a console that can be launched using python_console.launch. Here a scripts utils is provided which contains the definition of functions for transformations.

This folder contains:

• poses.json:
  • Description: A collection of TCP (tool center point) poses, referring to the world frame
  • Formant: "name_pose" : {position : {x,y,z}, orientation : {x,y,z,w}}.
  • Doc: Look here for more informations on Pose.
• scenes.json:
  • Description: A collection of Scene.
  • Formant1: "name_scene" : {name, color, position, orientation, type, dimensions}.
  • Formant2: "name_scene" : {name, color, position, orientation, type=mesh, file, scale}.
    • Specify the format (.stl) in the "file" field.
    • Check the position of the origin on cad programs.
    • Use the scale field if the file is exported to a unit other than meters.
  • Doc:
    • Click here for more informations on SolidPrimite.
    • Click here for more informations on ColorRGBA.
• meshes/: In this folder are contained the stl files

These files have been created to easily set up pick and place tasks. There is the template.cpp file with the skeletron to create your custom task.

Note: all launch files containe the "real_robot" parameter which, if set to false, disables the gripper_action_client allowing to work without the real robot.

• Use simulation.launch file for launch simulation with real_robot=False.
• Use controller.launch file for control the real robot with real_robot=True.

This node is a console for performing these simple tasks:

• gripper homing: homing the gripper

• gripper width: shows the gripper width. Remeber to use homing before to initiliaze.

• gripper 'width' 'speed': moves the gripper with the specified speed on the specified width. Default speed value is setted in launch file.

• grasp 'width' 'speed' 'force' 'epsilon_inner' 'epsilon_outer': performs the grasping. Default speed,force,epsilon values are setted in launch file.

• get joints: shows the joints values.

• get pose: shows the TCP(tool center point) pose, referring to the world frame.

• move joints 'j0' 'j1' 'j2' 'j3' 'j4' 'j5' 'j6': moves the arm joints in the specified configurations in Radian.

• move pose 'px' 'py' 'pz' 'ox' 'oy' 'oz' 'ow': moves the TCP in the specified pose.

• move linear 'px' 'py' 'pz' 'ox' 'oy' 'oz' 'ow': moves the TCP in the specified pose traveling in a straight line.

• ready: moves the arm in ready configuration.

• rel pos 'px' 'py' 'pz': moves the arm TCP along the x,y,z directions in Meters.

• rel rpy 'roll' 'pitch' 'yaw': rotates the arm TCP along the x,y,z axis in Degree.

• convert 'x' 'y' 'z' 'w': converts the quaternion in RPY (Radian).

• convert 'roll' 'pitch' 'yaw': converts the RPY (Radian) in quaternion.

• reset scene: resets scene.

• scene 'name': loads specified scene.

• save 'name': saves the current pose with the specified name. Default name is "last".

• goto 'name': moves the TCP in the saved pose named 'name'.

• set velocity 'vel': sets the arm max velocity scaling factor.

• quit: closes the node.

• help: shows the commands list.

Note: ',' are automatically ignored

This node execute the "pick and place" task. Two objects are moved in this example.

In the launch file you can modify the following parameters:

Parameters for startup:

• real_robot: if you want control the real robot set it to True to enable franka_gripper action client.
• gripper_homing: if you want to do the homing of the gripper on startup set it to True.

Parameters for gripper move:

• gripper_speed: sets the gripper speed.

Parameters for grasp:

• grasp_speed: sets the grasp speed.
• grasp_force: sets the grasp force.
• grasp_epsilon_inner: sets the grasp epsilon inner.
• grasp_epsilon_outer: sets the grasp epsilon outer.

Parameters for arm:

• arm_velocity_factor: sets the arm max velocity scaling factor.
• eef_step: sets the end-effector-step for linear motion.
• jump_threshould: sets the jump threshould for linear motion.

Parameters for pick and place:

• object_width_I: sets the grasp width for the I-th grasp
• pre_grasp_name_I: sets the name of the TCP pose saved for the I-th pre-grasp pose
• grasp_name_I: sets the name of the TCP pose saved for the I-th grasp pose
• post_grasp_name_I: sets the name of the TCP pose saved for the I-th post-grasp pose
• pre_place_name_I: sets the name of the TCP pose saved for the I-th pre-place pose
• place_name_I: sets the name of the TCP pose saved for the I-th place pose
• post_place_name_I: sets the name of the TCP pose saved for the I-th post-place pose Add this parameters for each task you want.

These nodes performs a simple teleoperation.

This node read stdin for sending a message on the topic "panda_controller/teleoperation".

Use Mode to switch mode and INCREASE and DESCREASE KEYS to change the value of DELTA.

In the launch file you can modify the following parameters:

• frequency: sets the frequency of the read and send cycle.
• start_delta_position: sets the start delta in position value.
• start_delta_orientation: sets the start delta in orientation value.
• start_delta_gripper: sets the start delta in gripper value.
• granularity_position: sets the granularity for position.
• granularity_orientation: sets the granularity for orientation.
• granularity_gripper: sets the granularity for gripper width. Then are available the parameters to modify the mapping of the commands.

This node read from the topic "panda_controller/teleoperation" a message for moving the arm, and moves it.

In the launch file you can modify the following parameters:

Parameters for startup:

• real_robot: if you want control the real robot set it to True to enable franka_gripper action client.
• gripper_homing: if you want to do the homing of the gripper on startup set it to True.

Parameters for gripper move:

• gripper_speed: sets the gripper speed.

Parameters for grasp:

• grasp_speed: sets the grasp speed.
• grasp_force: sets the grasp force.
• grasp_epsilon_inner: sets the grasp epsilon inner.
• grasp_epsilon_outer: sets the grasp epsilon outer.

Parameters for arm:

• arm_velocity_factor: sets the arm max velocity scaling factor.
• eef_step: sets the end-effector-step for linear motion.
• jump_threshould: sets the jump threshould for linear motion.

I used these extensions:

• c/c++ by microsoft
• c++ intellisense by austin
• cmake by twxs
• doxygen by christoph schlosser
• clang-format by xaver
• doxgen documentation by christoph schlosser
• python by microsoft
• git graph by mhutchie
• gruvbox mirror by adamsome
• vscode-icons by icons for visual studio code
• git graph by mhutchie

Are available the following tasks:

• doxygen: generates the documentation.
• format: formats all files.
• cloc: returns code statistics.

Enrico Sgarbanti @Envq.

This project is licensed under the GPL v3 License - see the LICENSE.md file for details.