In this lesson, we’re going to lay the foundation for the entire module by setting up a clean ROS2 workspace and integrating the xArm ROS2 stack, which will be the core robotic platform for all the applications we’ll develop later — from blind pick-and-place, to vision-based manipulation, to CNC-style toolpaths, all the way to full LLM-driven reasoning.

Before we jump into motion control, perception, and LLM planning, we need a reliable and well-organized development environment. This is exactly what today’s lesson is about.

What you will learn in this lesson

By the end of this lesson, you will be able to:

1. Create a new ROS2 workspace dedicated to LLM-powered robotics applications.
2. Clone the full xArm ROS2 ecosystem, including the robot description, MoveIt configuration, and control packages.
3. Build the workspace and verify that the robot can be launched successfully.

Why we start here

A solid foundation is critical when building advanced robotic applications.
Throughout this module, you will:

• Create custom packages
• Integrate MoveIt planners
• Add a simulated camera
• Build a perception pipeline
• Implement CNC trajectory generation
• And finally, connect everything to an LLM that generates motion commands and task plans in JSON

All of this depends on having a clean and stable robot environment from the beginning.
That’s why Lesson 1 focuses exclusively on preparing your workspace the right way.

What we are going to do step-by-step

1. Create a ROS2 workspace called xarm_ws.
2. Clone the main repositories:

• • xarm_ros2

1. Install dependencies using rosdep.
2. Build the workspace with colcon.
3. Source the environment.
4. Launch the robot and verify everything works.

Once these steps are complete, you’ll have a fully functional robotic base ready to be extended with custom applications.