MIT-REALM/franka_joint_velocity_controller

A joint-space velocity controller for the FR3

Joint-space velocity controller for Franka

Dockerized Franka ROS interface

This repository allows you to run libfranka and franka_ros in a Docker environment (complete with ROS Noetic, GUI access, and networking to the robot with the FCI_IP environment variable).

To build:

docker compose build

To run

docker compose run franka

Now you should have access to the full Franka ROS interface documented here.

For example, you can test the connection to the robot as follows:

1. Go to the Desk webapp and login.
2. Unlock the robot using the sidebar on the right of the screen.
3. Open the drop-down in the top right by clicking on the IP address, then click "Activate FCI" to enable the Franka Control Interface.
4. Run the docker container using docker compose run franka
5. Run the command roslaunch franka_example_controllers joint_impedance_example_controller.launch robot_ip:=$FCI_IP load_gripper:=true robot:=fr3

The robot should move in a circle, and an RViz window should open showing a matching 3D model of the robot.

Controlling the robot via ROS

In theory, the franka_ros Docker image provides a full installation of all ROS packages needed for full control of the arm. Unfortunately, "full control" means writing a C++ controller to run in real time. To support easier control of the arm, we provide a docker compose option that will launch a cartesian impedance controller that will track an equilibrium pose published as a ROS topic.

To run this controller:

1. Run docker compose up. This will launch 3 Docker containers:
   • franka_ros-roscore-1: runs roscore
   • franka_ros-cartesian_impedance_control-1: runs the impedance controller, RViz (for visualizing the state of the robot), and a GUI utility for modifying the impedance control stiffnesses of the robot.
   • franka_ros-main-1: runs a bash shell (this is where you should run your code).
2. In a new terminal tab, run docker attach franka_ros-main-1. This will give you a bash shell that can talk to the robot via ROS. From here, you can publish geometry_msgs/PoseStamped messages to the /cartesian_impedance_example_controller/equilibrium_pose topic to move the arm.

When you attach to the franka_ros-main-1 container, you should see that any files in the franka_realm folder have been mounted to the container. This folder is a ROS package, so any scripts you add there will be accessible from inside the container.

Adding new components

By default, the docker-compose.yml file launches 3 containers (called "services"):

• roscore: runs roscore.
• cartesian_impedance_control: runs the impedance controller to track an end effector pose.
• main: to provide a bash terminal for the user.

You can add additional services by adding them to the docker-compose.yml file:

services:
  
  ...

  my_new_service:
    image: franka_ros:latest  # which docker image to use
    build: .docker  # where that dockerfile is stored
    environment:
      - DISPLAY=${DISPLAY}  # allows GUI access
    volumes:
      - /tmp/.X11-unix:/tmp/.X11-unix  # allows GUI access
      - ./realm:/catkin_ws/src/realm  # allows access to the ROS package in the realm directory
    network_mode: host  # allows access to the host's network (and the robot)
    privileged: true  # needed for GUI + USB access
    stdin_open: true # allows you to attach to a shell in this container
    tty: true        # allows you to attach to a shell in this container
    depends_on:  # don't start until these have started
      - roscore
      - cartesian_impedance_control
    command: bash  # the command to run when starting this container.

The main things to edit here are changing which volumes get mounted (e.g. if you want to access different files) and the command to run when started (e.g. if you want to run some launch file from a package you've mounted). If you want to install additional dependencies in the docker image, you can modify .docker/Dockerfile.