Robotic Arms & Manipulation Course with ROS2 & MoveIt2

You build a working robotic arm across nine hands-on sections, each ending with something tangible you can show.

What you will learn and build:

• ROS2 from scratch: workspaces, nodes, topics, and launch files
• Communication and control: topics, services, and actions, with a focus on why arms rely heavily on actions for motion planning
• Modeling an arm: describe your own arm in URDF/xacro, add a gripper, and view it live in RViz moving joint by joint
• Kinematics: understand forward and inverse kinematics, tell the arm "put your gripper here" and watch it work out the joint angles
• MoveIt2 motion planning: generate a config with the Setup Assistant, plan and execute collision-free paths from code
• Collision-aware manipulation: teach the arm about obstacles, plan around them, and carry objects through cluttered scenes
• Pick and place: build the full grasp cycle (approach, grasp, lift, move, place) and simulate it end-to-end in Ignition Gazebo
• Real hardware: drive a physical arm through ros2_control, bring up a real gripper and camera, and run your pick-and-place routine on actual hardware
• AI perception and LLM control: build OpenCV and deep-learning perception nodes so the arm knows exactly where objects are in 3D space. Say "pick the blue one" and it does. Say "put the block in the box" and perception plus planning carry it out.

Capstone projects:

• Vision-guided pick-and-place
• Sorting and kitting station
• Language-commanded manipulation
• LeRobot imitation-learning arm

Who is this for?

Whether you are an engineering student, a working professional, or a builder ready to move past tutorials, this course takes you from zero ROS2 experience to deploying an intelligent robotic arm that plans, grasps, sees, and listens. Basic Python or C++ knowledge is all you need to start.