Serial Link Action Server

This ROS2 package provides a set of action servers for controlling serial link robot arms. It builds upon RobotLibrary C++ library which contains the fundamental algorithms for kinematics, dynamics, and control.

Note

This package is still under development. More features will be added.

Table of Contents

• Overview
  • Components
• Prerequisites
• Installation
• Usage
  • Launching the Nodes
  • Testing the Action Server
• Contributing
• License

Overview

Components

These are the current basic components that show how to coordinate the different modules to enable real-time control of a robot arm. The design is modular, and you can create your own action servers & action clients as needed. More actions & features will be added in future.

• include/ModelUpdater.h: This uses a pointer to a RobotLibrary::KinematicTree object which defines a robot model. It subscribes to a sensor_msgs/msg/JointState.msg topic and uses it to compute the forward kinematics and inverse dynamics based on the current state. The subscription topic name can be set as a constructor argument.
• src/velocity_control_server.cpp: This creates the RobotLibrary::KinematicTree model, a RobotLibrary::SerialKinematicControl controller, and links together the ModelUpdater node, as well as creating & advertising the TrackJointTrajectory and TrackCartesianTrajectory action servers. It publishes joint commands using a custom JointCommand.msg message. You can write your own ROS2 subscriber to subscribe and send these to your robot.
• src/demo_client.cpp: This communicates with the velocity_control_server to demonstrate how to perform joint & Cartesian control actions.
• src/mujoco_relay.cpp: This node works in conjunction with this MuJoCo interface 3D simulation. It subscribes to the JointCommand.msg topic from the action servers, and passes on a std_msgs::msg::Float64MultiArray.msg which is then used to control the simulated robot.
• launch/mujoco_velocity_mode.py: This contains various launch parameters, such as the URDF file for the robot model, control gains in the /config directory, subscription & publisher names, and launches the mujoco_relay node in conjunction with the MuJoCo simulation.

Prerequisites

• ROS2, obviousyl!
• RobotLibrary: This contains the fundamental algorithms for forward kinematics, inverse dynamics, and feedback control.
• Eigen: This is actually utilized by RobotLibrary, so if that is working then this is satisfied.
• MuJoCo Interface: An optional package. It is used to demonstrate this package with a 3D simulation.

Installation

1. Clone the repository:

   git clone https://github.com/Woolfrey/server_serial_link.git
   

2. Build the package:

   Navigate to your ROS2 workspace and build the package:

   cd your-ros2-workspace
   colcon build
   

3: Source the workspace:

```
source install/setup.bash
```

Usage

Launching the Nodes

There is a launch file designed to work with the mujoco_interface package.

You may modify/write your own to work with any robot.

Testing the Action Server

1. Launch the MuJoCo simulation. There are two options, velocity_mode.py, or torque_mode.py, but currently the action server is only working in velocity mode:

ros2 launch mujoco_interface velocity_mode.py

2. Launch the action server. It must match the control mode:

ros2 launch serial_link_action_server mujoco_velocity_mode.py

3. Run the client:

ros2 run serial_link_action_server demo_client <numberOfJoints>

Replace with the number of joints of the robot. Type options to see a list of commands.

Tip

The client is independent of the control mode! 👍

Contributing

Contributions are welcome! Please fork the repository and submit a pull request with your changes.

License

This project is licensed under the GNU General Public License v3.0. See the LICENSE file for details.