Boston Dynamics, a U.S. robot company under Hyundai Motor Group, has announced t

Boston Dynamics, a U.S. robot company under Hyundai Motor Group, has announced the retirement of its hydraulic humanoid robot Atlas.

It is predicted that Boston Dynamics will abandon the hydraulic drive method, which is expensive and is evaluated as an old-fashioned technology, and introduce a new electric drive humanoid robot.

With the humanoid war intensifying recently in the U.S. and China's Big Tech, such as Tesla, Figure AI, and Unity, Boston Dynamics is also joining the competition. For Boston Dynamics, which has promised to be listed by the first half of 2025, the launch of new humanoid products could determine corporate valuation.

Since we first started with RoMeLa's "Electric" ARTEMIS humanoid two years ago, there have been a lot of humanoid robots following suit over the past year. Boston Dynamics' electric Atlas humanoid (secret code "ATLAS-E") comes among a number of new electric humanoid robots, including Tesla's Optimus, Figure 01, 1X's NEO, Apptronic's Apollo, Unitree's H1, Fourier Intelligence's GR-1, China-made fake robots and many more, including Mentee Robotics' MenteeBot.

[Hydratic vs. Full electric, and start-up ARTEMIS]
When we first unveiled RoMeLa's "electric" ARTEMIS humanoid two years ago, Artemis showed the world the potential of an electric humanoid for the first time with the best walking performance. And so many robots followed suit, but none of them could outperform Artemis' bipedal technology. So far...

When we first introduced the "electric" ARTEMIS humanoid, many people compared it to the "hydraulic" Atlas humanoid from Boston Dynamics. Although we have done a lot of work to explain to people why electric is far superior to hydraulic, it may not be long before we see the true face of Boston Dynamics' electric Atlas soon. Let's look forward to some surprising results again.

Using hydraulic power, hydraulic ATLAS, which retired yesterday, can produce enough power and power for ARTEMIS to match. However, unlike electric-powered ARTEMIS, it is noisy, complex, mesh, and has low energy efficiency, so there are many problems to be solved compared to electric (especially indoors) in practical terms. As far as I know, ATLAS is too powerful to match electric ARTEMIS, but ARTEMIS is faster. So Boston Dynamics also started researching electric-powered humanoids (relatively late), and finally today, they introduced a new electric-powered Atlas. However, surprisingly, today's new electric ATLAS is said to be stronger than hydraulic. It's also magic-like Boston Dynamics' amazing technology.

[But why now?]
So how has the humanoid field suddenly become so popular that the only thing that can be used in the past year is the expression "crazy"? I don't think this is the result of the subtle interwoven and exquisite timing of new technologies, capital, marketing, the desire for physical data, etc.

* The emergence of high-torque, compatible electric actuators capable of force control, such as RoMeLa's BEAR, rather than servomotor with stiff position control
* Control of new approaches to walking with MPC control and Sim2Real learning, and a deeper fundamental understanding of bipedalism
* The expectation that Machine Learning (which we call 'artificial intelligence') will solve problems that actually make humanoid robots useful, such as manipulation
* Tesla's Elon Musk's "Marketing" (some call it "vision"), which first hit a humanoid robot as a real business...)
* The Chinese government's huge investment in humanoids at the national level, and Chinese humanoid startups popping up like a mushroom as a result
* "Wake up" (or "fantasy") about the importance of humanoid robots as a tool to get the "physical" data needed to train humanoid robots, rather than "informational" data (text, image, video) that requires generative AI such as ChatGPT, DALL-E, and SORA
* As a result, data-thirsty U.S. VCs' huge investment in U.S. startups has created a strong capital
* And efforts by tech giants that already have enormous capital to develop AI technologies and apply them to robots (NVIDIA)
* And our RoMeLa research center and several other university robot research institutes in Accademia research the state-of-the-art technologies and cultivate and produce the high-quality ingenes that these companies need.

And so on… And I think that's why the timing of all these factors is so exquisite. It's actually a very exciting time for everyone, not just for the humanoid researcher! But....

[Unmanned car vs humanoid robot]
As you may know, the autonomous vehicle industry has been booming since the 2004/2007 DARPA Grand Challenge and Urban Challenge. 🚗👍 (Our RoMeLa lab is ranked third in the world!) On the flip side, since the 2015 DARPA Robotics Challenge, a humanoid robot competition, we've had a run-off -- we've had a problem with Virginia Tech…) We expected the humanoid robot industry to be huge, but the disappointment with the robots' performance at the competition has resulted in a huge decrease in interest and funding in the humanoid sector. 🤖👎

By the way…
Now it's all the other way around!
Despite the huge investment and high expectations for driverless cars, investment in driverless cars has dwindled and many companies have begun to shut down one after another. 🚗👎 And conversely, investment and expectations for humanoid robots are now sky high for the reasons mentioned above. 🤖👍

What can we learn here?
I hope you think about it slowly.

[Where do we/I go from here?]
Our lab RoMeLa, including DARwIn, CHARLI, THOR, THOR-RD, SAFFiR, BRUCE, ARTEMIS (and a secret humanoid to be released in July), has produced perhaps the world's largest variety of humanoids and has always surprised the world with new approaches. It has produced a lot of innovations and amazing results, including new BEAR-powered devices, model base MPC controls, new mechanical structures, different locomotion configurations, the available automobiles, the dissemination through open sources, and so on. So, RoMeLa has been doing humanoid research for the last 20 years, and no one is against being one of the most outstanding laboratories in the world in this field. But where should the academic community be now in the field of humanoids that is growing mainly in companies?

Our lab, RoMeLa, unlike all of the companies mentioned above, is a school lab in academyia. Unlike companies with big capital and manpower, it's a relatively small school lab that I and my graduate students work on together. It's not like you feel confused and crisis-free between winning and releasing the title of "world's first" with companies around the world in this field that's developing so fast that you can't get your head around it. But we're excited and excited about new challenges again. That's because our ultimate goal is unwavering. We can take on challenges that are not afraid of failure, develop high-risk, hard-working, long-term research that companies can't. And we've already begun our work in the field of "new" applications that no other company can imagine. We're always flexible in different situations, but our ultimate goal is unwavering. A combination of infinite imagination and creativity, burning passion and fun, Romella always has the belief and belief that we're going to develop robot technologies that benefit humans. So I have very high expectations for the future.
By the way…

[Cautious optimism]
If expectations are high, disappointment is high?
Or,
The future is bright when you have high expectations?

Expectations for humanoid robots are sky-high. And once a week, amazing robots are released and new videos are posted. Investor funding is amazing on a scale, and leaders in this area offer a commitment/vision for our near future with humanoid robots. We promise that advances in AI will solve all the problems, and we have incredible intermediate results. But instead of taking all the videos, the stories we hear and the news in the media as it is, we need to analyze and accept it. While academic announcements are acceptable as they are, tech companies are not always like that. (There's a reason why I'm so confident.) You shouldn't be trailed, but you shouldn't fall too quickly or go the wrong way. I think that's a very important attitude we need to have in this fast-changing world.