LeRobot ROS

This repository provides a generic ROS 2 interface for the LeRobot framework. It acts as a lightweight wrapper to connect any ros2_control or MoveIt compatible robot arm with the LeRobot ecosystem.

A gamepad teleoperator for 6-DoF end-effector control and a keyboard teleoperator for joint position control is also provided.

Supported control modes:

• Joint position with ros2_control
  • Using joint_trajectory_controller
  • Using position_controllers
• End-effector velocity with MoveIt 2
  • Using Moveit Servo
• Gripper control with ros2_control
  • Using joint_trajectory_controller
  • Using Gripper Action Controller

Video Demo

Prerequisites

Software Requirements

Before getting started, ensure you have the following installed:

• ROS 2 Jazzy - This repo is only tested on Jazzy.
• ros2_control
• If end-effector control is desired, then MoveIt2 needs to be installed

Quickstart with Simulated SO-101

Below steps will allow you to perform keyboard teleoperation of a simulated SO-101 arm using Lerobot.

---

First, setup LeRobot and lerobot-ros in a virtual environment. Note that the Python version of the virtualenv must be compatible with your ROS version. For ROS 2 Jazzy, we use Python 3.12.

# Create and activate virtual env
conda create -y -n lerobot-ros python=3.12
conda activate lerobot-ros

# Install lerobot
git clone https://github.com/huggingface/lerobot
pip install -e lerobot

# Install lerobot-ros packages
git clone https://github.com/ycheng517/lerobot-ros
pip install -e lerobot-ros/lerobot_robot_ros
pip install -e lerobot-ros/lerobot_teleoperator_devices

Then, setup the Simulated SO-101 by following instructions in: https://github.com/Pavankv92/lerobot_ws

Finally, to run all programs:

# In terminal 1, run the Gazebo simulation
ros2 launch lerobot_description so101_gazebo.launch.py

# In terminal 2, load the ros2 controllers and run MoveIt
ros2 launch lerobot_controller so101_controller.launch.py && \
  ros2 launch lerobot_moveit so101_moveit.launch.py

# In terminal 3, run lerobot with the ROS version of so101 and keyboard teleop
cd <YOUR lerobot-ros DIRECTORY>
lerobot-teleoperate \
  --robot.type=so101_ros \
  --robot.id=my_awesome_follower_arm \
  --teleop.type=keyboard_joint \
  --teleop.id=my_awesome_leader_arm \
  --display_data=true

Once you have teleoperation working, you can use all standard LeRobot features as usual.

Robot Integration Guide

This section describes how to integrate other ROS-based robots with Lerobot.

Arm Control Modes

Currently the repo supports the following arm control modes:

Option 1: Joint Position Control

This option uses position_controllers in ros2_control. It requires the robot to have:

• position_controllers/JointGroupPositionController for the robot arm joints
• joint_state_broadcaster/JointStateBroadcaster for joint state feedback

This option is enabled by setting action_type to ActionType.JOINT_POSITION in robot config.

Option 2: Joint Trajectory Control

This option uses joint_trajectory_controller in ros2_control. It requires the robot to have:

• joint_trajectory_controller/JointTrajectoryController for the robot arm joints
• joint_state_broadcaster/JointStateBroadcaster for joint state feedback

This option is enabled by setting action_type to ActionType.JOINT_TRAJECTORY in robot config.

Option 3: End-Effector Control

This option uses Moveit Servo in MoveIt. It requires the robot to have:

• The moveit_servo node for real-time end-effector control
• joint_trajectory_controller/JointTrajectoryController for robot arm control
• joint_state_broadcaster/JointStateBroadcaster for joint state feedback

This option is enabled by setting action_type to ActionType.CARTESIAN_VELOCITY in robot config. See: ar4_ros_driver for an example of using moveit_servo.

Gripper Control Modes

The repo supports two gripper control modes that can be configured via the gripper_action_type setting:

Trajectory Control (GripperActionType.TRAJECTORY)

• Uses JointTrajectoryController from ros2_control
• Publishes JointTrajectory messages to /gripper_controller/joint_trajectory

Action Control (GripperActionType.ACTION)

• Uses GripperActionController from ros2_control
• Sends action goals to /gripper_controller/gripper_cmd
• Provides feedback on whether the gripper reached its target position

Code Changes to Lerobot-ros

Extend the ROS2Robot class in robot.py. This class can be a simple pass-through. It's just is needed to satisfy lerobot device discovery requirements.

class MyRobot(ROS2Robot):
  pass

Then, create a config class for your robot by sub-classing ROS2Config in config.py. The name of this class must be the same as your robot class, suffixed by Config. You may override joint names, gripper configurations, and other parameters as needed. An example config class for joint velocity control may look like this:

from dataclasses import dataclass, field
from lerobot.common.robots.config import RobotConfig
from lerobot.common.robots.config import ROS2Config, ROS2InterfaceConfig

@RobotConfig.register_subclass("my_ros2_robot")
@dataclass
class MyRobotConfig(ROS2Config):
    action_type: ActionType = ActionType.JOINT_POSITION

    ros2_interface: ROS2InterfaceConfig = field(
        default_factory=lambda: ROS2InterfaceConfig(
            base_link="base_link",
            arm_joint_names=[
                "joint_1",
                "joint_2",
                "joint_3",
                "joint_4",
                "joint_5",
                "joint_6",
            ],
            gripper_joint_name="gripper_joint",
            gripper_open_position=0.0,
            gripper_close_position=1.0,
            max_linear_velocity=0.05,  # m/s
            max_angular_velocity=0.25,  # rad/s
        )
    )