Building a MicroDuck, Part 1: The Plan

MicroDuck is a small biped robot: two legs, a neck and a head, and a walk that was learned in a computer simulation before it ever touched a real floor. OpenELAB wrote a build guide for it, and there is a matching build tutorial on GitHub. I’m going to build it, and you can follow along, one small step at a time.

No robotics background? No problem. You don’t need to know what a servo or an IMU is yet. I’ll explain every new word the first time it shows up, tell you what each step does and why, and show you how to check that it worked.

Who Is This Series For?

• Beginners. If you can follow a recipe, you can follow this series.
• Younger builders. Go for it, but do the soldering and the battery steps together with an adult.
• People who want to see the real thing. I’ll show the mistakes, the real costs and the parts that didn’t work the first time.

How Every Part Works

1. What you’ll do in this part, and what you need.
2. Safety first, before any step that needs it.
3. Numbered steps. For each one: what to do, what’s happening, and why.
4. A check after each step: “Good result” and “Bad result”.
5. If it didn’t work: the common problems and how to fix them.
6. Words explained at the end of each part.

What Is MicroDuck?

MicroDuck has 14 servos in its legs, neck and head. The mechanical design has room for a 15th servo that moves a mouth, but the walking does not depend on it. A Raspberry Pi Zero 2 W is the brain. It reads a motion sensor and a gamepad, and runs a walking model stored in a file called walk.onnx.

The model was trained with reinforcement learning: a computer practises walking millions of times in a virtual world, gets a score for each try, and slowly gets better. The result is a “policy” that looks at what the robot senses and decides how every joint should move. The guide’s main advice is to run the supplied model first before thinking about training your own.

Which MicroDuck Is This?

There are two projects with almost the same name, and they are not the same robot.

I read Pollen’s design documents. Their software only supports the Radxa board and their own motion-sensor board, and says the Raspberry Pi is dropped. So their software will not run on this build as it is. I use the software from the OpenELAB build guide instead, and I’ll keep Pollen’s simulator in mind for Part 9.

What Will It Cost?

I would rather leave the total open than guess. Part 2 has the full shopping list, and I’ll fill in the prices there as I find them.

What I Want to Achieve

1. A robot that stands and walks with the ready-made walk.onnx model.
2. Driving it with a Bluetooth gamepad, without a laptop.
3. If that works: add some character, and look at training my own walk.

What I Don’t Know Yet

• Where the print files come from. The tutorial says to print from a cad folder, but I can’t find that folder in the repository. OpenELAB shows a matched set of parts. I’ll find out in Part 2 before I order anything.
• The exact battery. The tutorial says a ready-made 6 V rechargeable pack that can deliver enough peak current. It doesn’t name one.
• How the servos are configured without a U2D2. The tutorial lists a U2D2 as a setup tool, OpenELAB calls it optional.
• How well the walk works on my robot. A model fits one robot. If mine differs a little, it may fall over.

The Plan in Nine Parts

Each part is published when the step is finished, so you’ll see my mistakes too, and the plan may change along the way.

Words Explained

• Biped: a robot that walks on two legs.
• Servo: a small motor that turns to an exact position and holds it.
• Dynamixel: a brand of servos that you chain together on one cable.
• Raspberry Pi: a tiny, cheap computer. This one is the robot’s brain.
• IMU: a motion sensor that measures tilt and movement.
• Reinforcement learning: teaching a computer by trial and error.
• Policy / ONNX: the trained walking skill, stored in a standard file format.
• Open source: the designs and code are public.

Conclusion: A Small Duck with a Big Idea

A robot that learned to walk in a simulator is within reach of a hobbyist, as long as I’m honest about what I don’t know yet. Next up: the shopping list.

Small duck, big steps. Happy hacking!