Thread by @chubicki

Elon Musk revealed the #TeslaBot today and I have #robotics thoughts.

https://twitter.com/ElectrekCo/status/1576019791193657344

🧵

Am I blown away? No. Am I laughing? No.

First, the team did a good job. They came a long way in about a year(?), going from zero-to-robot from the ground up.

Also, doing a live demo without a tether (safety catch rope) is braver than people know.

Let's talk walking.

I told my lab today that I expected it to walk on stage, and it did, but I was a bit surprised how they did it.

It *seems* to use a method called Zero-Moment Point to maintain balance. It's been used in various forms since the 90's.

You see ZMP in the bent knees and how it shifts its weight over to its next foot before taking a step

It's pretty safe, but not mindblowing in 2022

Also, people don't walk quite like this. We are more efficient- we stick our foot out, fall, catch, repeat

People are justifiably comparing it to Boston Dynamics' ATLAS humanoid, but that's not the best comparison.

1. While BD sells their 4-legged Spot, it doesn't aim to make ATLAS in volume.
2. BD humanoids are designed for awesome demos, not practical tasks.

BD designs ATLAS with high-powered hydraulic actuators, instead of electric motors like Tesla, which enables their stunts.

Hydraulics are a tradeoff because they are power hungry. It's why ATLAS demos last for minutes and not hours -- battery drain.

A better corporate comparison would be Agility Robotics. They've always used EM motors. After 18 months they had their first walking prototype, Cassie

You'll note it walks completely differently. Cassie balances dynamically, catching itself every step

Agility took another 2 years to make Digit, which is a humanoid with arms:

Agility is now on Rev 4 of Digit which has a full sensor package and has sold dozens of units.

Full disclosure: They spun off from my PhD lab and I contributed to their early work

Also in the early 2010's, the DARPA Robotics Challenge (DRC) gave teams ~18 months to make their first round of humanoid robots for a major competition.

So Tesla's year-ish to a humanoid is impressive by those standards, but those DRC teams did well too.

The dancing, from a controls standpoint, is nothing crazy. It's something you can do with inverse kinematics -- a standard method for computing robot poses.

Here's Perdue students using a Rainbow Robotics HUBO to do a dance.

There's a portion showing off the vision, manipulation, and "smarts." It's hard for me to glean the methods here because it's a cut video, so I don't have much to say. I don't know how much is pre-canned vs. online planning.

I will also let everyone in on a secret about humanoids:

It's all about reliability. How often does it fall down? You can't tell from a cool video -- or even a live demo. Boston Dynamics, to its credit, has been upfront in showing when its robots fail.

He mentioned a $20K price point, which would be a big deal. The cheapest human-sized humanoids I'm aware of are in the $150K range (a long way from the olden days of $1M price tags).

.... but I'll believe the price when I can actually buy one.

My take:
1. Impressed by the short turnaround to build a new humanoid robot and show it live. Hats off to the team.
2. The shown capability seems standard (but not mind-blowing) for humanoids.
3. Zero idea on the reliability.
4. I'll believe the price when my lab buys one.

P.S. I can only *imagine* the stress on the demo team leading up to this. There's no aesthetic covering on the chest of the walking model and I just know a boss was saying "Why is it naked!? Put the cover on!" and the engineers were like... WE'RE WORRIED ABOUT VENTILATION AHHHHH!