franka-emika-guide

User guide for the Franka Emika robots at TU Delft

For more info: https://frankaemika.github.io/docs/installation_linux.html

Guide to getting started with the Franka Emika Robots's

This guide is to help people get up and running with the Franka Emika Panda's for the first time and as a reference for people already familiar with using the Panda's in the lab.

For someone using the Panda for the first time, the guide should be read sequentially.

libfranka

libfranka is the software package on the PC connected directly to the Panda controller that supports the communication interface necessary to send torque commands and receive information such as position, velocity, mass matrix etc.

This package should already be installed on the PC's in the lab that are used to run the controllers. It is not necessary to install this on your own PC, unless you are running the Panda in the Gazebo environment.

Installing libfranka libraray

On the computers in the lab, this step should be skipped. The computers should have libfranka installed. Don't reinstall it or upgrade/downgrade it, as other students code might rely on this version.

If you do need to install the libfranka library, there is a guide from Franka Emika, which is summarised below for convenience. If this guide doesn't work, consult the official guide at, https://frankaemika.github.io/docs/installation_linux.html , as these instructions might become outdated.

• Open a terminal by pressing ctrl+alt+t

• In case you already have some versions of libfranka installed, remove them to avoid conflicts with:

sudo apt remove "*libfranka*"
sudo apt autoremove

For the Panda: Type the following commands to generate and build libfranka

For the Panda:

cd
sudo apt install build-essential cmake git libpoco-dev libeigen3-dev
git clone --recursive https://github.com/frankaemika/libfranka
cd libfranka
mkdir build
cd build
cmake -DCMAKE_BUILD_TYPE=Release ..
cmake --build .

For the fr3:

cd
sudo apt install build-essential cmake git libpoco-dev libeigen3-dev
git clone --recursive https://github.com/frankaemika/libfranka --branch 0.13.3 # only for FR3
cd libfrankaecursive https://github.com/frankaemika/libfranka
cd libfranka
mkdir build
cd build
cmake -DCMAKE_BUILD_TYPE=Release ..
cmake --build .

This last command may take several minutes to compile the code. Also, the version (in this case 0.13.3) is constantly getting updated by Franka Emika. So check there install instructions (warning, these don't get updated frequently enough and can be behind what is needed) and github for the most up to date instructions and versions of libfranka.

Setting up the franka_humun_friendly_controllers

franka_ros is a ROS support package supplied by Franka Emika (https://frankaemika.github.io/docs/franka_ros.html). It contains supporting materials such as CAD files, urdf's and templates for controllers. The human franka_human_friendly_controllers package adds more functionality on top of what is given in the base franka_ros package, especially features to help researchers that use impedance based control and need human friendly interactions.

On the controller Desktop computer in the lab, set up a catkin workspace for yourself at ~/Documents/your_name_ws/, .e.g. ~/Documents/giovanni_ws/.

cd to your workspace, download and install the franka_ros package

cd ~/Documents/your_name_ws/
mkdir src
cd src
git clone --recursive https://github.com/frankaemika/franka_ros 

Then install the franka_human_friendly_controllers within the franka_ros folders

cd src/franka_ros
git clone https://github.com/franzesegiovanni/franka_human_friendly_controllers.git

Then build the packages, sourcing the compatable version of libranka, depending the robot model, panda, or fr3:

Panda version

catkin_make -DMAKE_BUILD_TYPE=Release -DFranka_DIR:PATH=~/libfranka_panda/build

or

catkin build -DMAKE_BUILD_TYPE=Release -DFranka_DIR:PATH=~/libfranka_panda/build

fr3 robot version

catkin_make -DMAKE_BUILD_TYPE=Release -DFranka_DIR:PATH=~/libfranka_fr3/build

or

catkin build -DMAKE_BUILD_TYPE=Release -DFranka_DIR:PATH=~/libfranka_fr3/build

Running the controllers on the Robot

• Switch on the power switch of the controller
• On the PC, set the ethernet too "connect to panda", i.e. manually set the IP address to 172.16.0.1
• In a web browser (.e.g.firefox), type in the IP address of the robot you are using, a Franka Emika browser application screen should appear
  • Tip: If the browser application is not found, try using the ping command from the command line e.g. ping 172.16.0.2. There are multiple Panda's in the lab with multiple IP address, this can be quick way to check which on is connected (we probably should write the ip address on each robot!)
  • Tip: Check that both emergency power switches are not engaged (The fr3's only have one, the Panda has an extra emergency between stop between the controller and the power socket.
• In the browser application, unlock the robot joints.
• In the menu (three bars at the right top corner), activate the fci
• In the terminal, source the ROS packages you built source ~/Documents/your_name_ws/devel/setup.bash
• Then, to you run the controller on the robot run:

roslaunch franka_human_friendly_controllers <controller_name>.launch robot_ip:=<ROBOT_IP> load_gripper:=True

for example, the Cartesian impedance controller with the robot that has an IP address of 172.16.0.2:

roslaunch franka_human_friendly_controllers cartesian_variable_impedance_controller.launch robot_ip:=172.16.0.2 load_gripper:=True

• If you need to use the gripper, in a different terminal, run (don't forget to source again):

rosrun franka_human_friendly_controllers franka_gripper_online

Connecting to the Robot with your own computer

1. Connect your PC/laptop to the network that ROS_MASTER is running on .i.e. connect your PC/laptop to the network switch with the ethernet cable
2. Create a new wired network and in IPv4 set Manual and put a new ip for your computer, in Ubuntu settings <PC/laptop>=A.B.C.F where F is different from the IP address of the computer connected to the robot controller, <controller_computer_ip> (which is ussually 172.16.0.1 in the lab) or the <robot_ip> (which is in the range of 172.16.0.2 - 172.16.0.5, for the four Franka Emika robots we have in the lab). Netmask is the same 255.255.255.0. Save the network.

3. Add this to your bash file (gedit ~/.bashrc):

export ROS_MASTER_URI=http://<controller_computer_ip>:11311
export ROS_IP=<PC/laptop>
export ROS_HOSTNAME=<PC/laptop>

For example

export ROS_MASTER_URI=http://172.16.0.1:11311
export ROS_IP=172.16.0.25
export ROS_HOSTNAME=172.16.0.25

Note to switch the ROS_MASTER_URI back to the default settings, add the line

export ROS_MASTER_URI=http://localhost:11311

to the ~/.bashrc file and comment out the previous export lines. Then re source the ~/.bashrc file

4. source the ~/.bashrc file

source ~/.bashrc

5. source ros source /opt/ros/<ros_version>/setup.bash

6. Test the data_streaming with rostopic list

Running the robot in Gazebo

This section is copied from https://github.com/franzesegiovanni/franka_human_friendly_controllers/, if there is any problems with using the instructions below to use Gazebo, please check there for a more up to date version and let us know of the problem.

The first time you run it, set up Gazebo with python3 setup_gazebo.py (in the 'franka_human_friendly_controllers' package).

To lunch the cartesian impedance controller in simulation:

roslaunch franka_gazebo panda.launch x:=-0.5 world:=$(rospack find franka_gazebo)/world/stone.sdf controller:=cartesian_variable_impedance_controller rviz:=true

To kill gazebo run: killall -9 gazebo & killall -9 gzserver & killall -9 gzclient

Bimanual Setup

Learning Resources

Some resources to get up to speed with the prequisites for using the Franka Emika robots and ROS in general. A lot of these are hijacked from https://www.ipb.uni-bonn.de/onboarding/. Also, these haven't been tested, so if you do use one of these, please let us know if it was useful or not, or if know of other good resources you have actually used, so we can curate the list over time.

ROS

Lecture series: https://www.youtube.com/playlist?list=PLE-BQwvVGf8HOvwXPgtDfWoxd4Cc6ghiP

Official ROS tutorials: https://wiki.ros.org/ROS/Tutorials

Linux/Ubuntu

Recommendations are very welcome for this!!

Git

A beginners guide: https://rogerdudler.github.io/git-guide/

More advanced course: https://missing.csail.mit.edu/2020/version-control/

Interactive tutorial: https://learngitbranching.js.org/

Another fundamentals tutorial: https://www.youtube.com/watch?v=HVsySz-h9r4

A very hand guide for when you're in trouble: https://ohshitgit.com/

Python

Python Intro from Bonn University: https://www.youtube.com/playlist?list=PLgnQpQtFTOGRKTPxhXZ68kKAQtsy76s0S

C++

Two courses from Bonn University: https://www.youtube.com/playlist?list=PLgnQpQtFTOGR50iIOtO36nK6aNPtVq98C https://www.ipb.uni-bonn.de/cpp-course-2020/