Why this unit exists: The most common reason people abandon a robotics tutorial is not difficulty — it's getting stuck mid-session because they are missing a cable, their Python version is wrong, or they don't know who to ask. This unit exists so that never happens to you.

What You Will Build

By the end of this path you will have a working pick-and-place policy running live on your OpenArm. You will teleoperate the arm to collect 50 demonstrations, train an ACT imitation learning policy on those demonstrations, and deploy it. The arm will complete the task autonomously — no human input required.

This is not a toy example. It is the same workflow used by research labs collecting robot learning data at scale. Once you have done it once, you can apply it to any task.

Hardware Checklist

Before starting Unit 1, verify you have every item below in hand.

  • OpenArm 101 — the 6-DOF robot arm. Order here if you haven't already. Allow 3–5 business days for delivery.
  • USB-C to USB-A cable — for initial connection to your computer. A short (0.5–1m) cable is easier to manage.
  • Power supply — included in the box. Verify it arrived undamaged. The arm requires 24V DC.
  • Clear workspace — at least 60cm × 60cm flat surface. The arm's reach is ~50cm. Remove anything fragile within reach.
  • Mounting option — a clamp mount or bolt-down plate. The arm should not be freestanding during operation. Both options come in the box.
  • Teleoperation device — either a spare OpenArm as a leader arm (recommended) or a VR controller. Details in Unit 3.

No physical hardware? You can complete most of this path in simulation. See the simulation setup guide before Unit 1.

Software Checklist

These are the software requirements. Confirm each item before moving on.

  • Ubuntu 22.04 or 24.04 — required for ROS 2 and SocketCAN. A VM works; dual-boot is better for real-time performance. macOS and Windows are not supported for hardware operation.
  • Python 3.10 or higher — run python3 --version to check. Install via sudo apt install python3.10 if needed.
  • pip and venv — run pip3 --version. Install via sudo apt install python3-pip python3-venv.
  • Git — run git --version. Install via sudo apt install git.
  • ~15 GB free disk space — for ROS 2, the SDK, LeRobot, and your dataset. Training datasets can grow large.
  • GPU (optional but recommended for Unit 5) — an NVIDIA GPU with 8GB+ VRAM speeds up ACT training from ~3h to ~45min. Training on CPU is possible but slow. Cloud GPU options are covered in Unit 5.

Time Estimates

These are realistic estimates, not optimistic ones. Factor in troubleshooting time.

Unit What You Do Time
0This orientation20 min
1Hardware setup, first motion~2 h
2SDK install, Python connection~1.5 h
3Teleoperation session~2 h
4Record 50 demo dataset~2 h
5Train ACT policy~3 h
6Deploy and evaluate~1.5 h
Total ~12 h 20 min

You do not need to complete the path in one sitting. Units 1–4 should be done across at most 2 sessions (hardware setup is easier when you pick up where you left off). Unit 5 training can run unattended — start it before sleep.

How to Get Help

You will not be alone. If you get stuck, use these resources in this order:

  1. Check the completion check at the bottom of whatever unit you're in — it often tells you exactly what to look for.
  2. Read the troubleshooting section in Unit 2 and Unit 3 — they cover the 80% case for connection and teleoperation issues.
  3. Post in the OpenArm forum thread — include your Ubuntu version, the exact error message, and which unit you're in.
  4. Join the SVRC Discord in #openarm-path — faster response during daytime hours in PST.

Please do not skip ahead to the next unit while blocked. The path is sequential by design — being stuck at Unit 2 usually means Unit 1 is not fully complete.

Orientation Complete When...

You have checked every item in the hardware and software checklists above, you know your path to the forum if you get stuck, and you have set aside your first 2-hour session for Unit 1.