Mudras for robotic hand alphabet

The extraordinarily long “Humanoid Robot” thread shows the passionate disagreement between those who believe that the world is better served by humanoid-shaped robots vs. those who believe that dedicated non-humanoid robots would be better.

But I’m sure that we can all agree that medically-prescribed prosthetic hands would be better served by robotic hands with the flexibility and dexterity of real human hands. This is a real challenge. I recall a little girl in my first grade class who was missing both her hands and used prosthetics with grasping devices that resembled needle-nose pliers. (1958)

The key insight for me is that hand motions can be described by an “alphabet” of motions that can be combined into the majority of actual hand movements. Imagine that! It’s the leap from a robot trying to mimic human hand motions (analagous to Chinese ideograms) and a robot programmed with a couple of dozen movements which can be combined into hundreds of potential movements (analagous to an alphabet).

Cool!

Indian researchers have extracted the specific hand motions practiced in the’ Indian classical dance form known as Bharatanatyam.

Reconstructing hand gestures with synergies extracted from dance movements

Scientific Reports volume 15, Article number: 41670 (2025)

….

The mudras are represented as fundamental syllables of motion. These syllables were then employed to reconstruct 75 diverse hand gestures, including American Sign Language (ASL) postures, a dataset of natural hand grasps and traditional mudras. Comparative analysis between mudra-derived synergies achieved superior reconstruction accuracy (95.78% for natural grasps and 92.99% for mudras) compared to synergies derived from natural grasps (88.92% for natural grasps and 82.51% for mudras). The results suggest that the structured and intentional nature of Bharatanatyam mudras leads to much stronger representation of syllables of movements that have superior generalizability and precision. Additionally, the reconstructed gestures were successfully mapped onto Mitra, a humanoid robot with five degree of freedom hand using a continuous joint-mapping approach. This research highlights the potential of dance inspired structured learning in enhancing dexterity, rehabilitation, and motor control, paving the way for more efficient gesture-based interaction models in robotics, prosthetics and rehabilitation. … [end quote]

Here is a video of some mudras. Try them out to see how agile your own hands are!

Here is a video of Bharatanatyam dance, a hymn to the Hindu god Shiva (who is often portrayed with multiple arms). Note the hand gestures (mudras) which are part of telling the story.

I have danced continuously since my mother enrolled me in a ballet class at age 4. I continue to dance almost daily and I plan to continue. A study on Parkinson’s disease revealed that an eight-week integrated dance therapy led to significant improvements in both cognition and quality of life. (And I’m pretty sure that these patients weren’t dancing Zumba with its fast, complex footwork.)

Back to robots…whether humanoid, parts of humans or even non-humanoid, I think that the designers of robots will benefit from the concept of a movement alphabet.

Wendy

5 Likes

Super interesting study on the mudras, thanks for sharing! Ran across this why surfing -

1 Like

Does it have to be either or? Not so. There is room for both. Dedicated non-humanoid robots have been with use for centuries and are of no interest to me as investments. Humanoid-shaped robots are the new kids in town and what METaR should be focusing on/ Which companies can monetize them? That is a thread I would live to follow.

The Captain

3 Likes

Tesla’s latest robot has 22 degrees of freedom (DoF) in the hands, provided by 17 actuators. A human hand has 27 DoF.

The next iteration has 50 actuators and may well exceed the dexterity of the human hand.

intercst

3 Likes

I bet that British girl has fun on Halloween.

intercst

2 Likes

Well … non-humanoid robots are also likely to do some exciting things as well going forward due to the same advances in computer-based control we are seeing for humanoid robots. True, a non-humanoid robot designed to occupy a fixed place in an assembly line may often not need the smarts of a mobile humanoid one, but examples like the ones of dealing with meat processing may be the exception. I expect some synergy with hand/manipulator developments as well.

OK, name one worth investigating.

I can name a humanoid one being developed right now and that has known jobs in factories.

The Captain

Anduril has some interesting stuff - YFQ-44A Fury. If you consider AI drones to be ‘robots’. I do.
Anduril is currently private, so, no direct investing opportunities.

:thinking:
ralph

Elevate Robotics is a spinoff of Apptronik. They’re developing some non-humanoid robots. Dexterity has a superhumanoid robot that is geared towards logistics -

2 Likes

Boston Dynamics is doing some pretty interesting stuff in locomotion …

3 Likes

One of the more significant implementations of new robotics has been Amazon’s integration of robots into their warehouses. They now have more than a million robots that are doing a lot of the logistical work of moving items and packages through the distribution chain. These aren’t humanoid robots pushing carts around. They’re Kiva robots, basically large “hockey puck” style flat wheeled robots that can move a half-ton of material around. The following article has a pretty good description of how they have significantly changed Amazon’s operations (the point of the article is mostly to advocate labor and unions at Amazon, but the description of how they’ve implemented the Kivas is very thorough):

In much of the video coverage of Amazon’s robots, you tend to see stock footage of Digit, the humanoid robot that Amazon has partnered with Agility Robotics to develop. But when Amazon talks about its robots, it does not mean humanoid service robots. It mainly means its Kivas: the Roomba-like robots that move things around its warehouse floors. The most widely deployed Kiva is the Hercules, which carries giant stacks of inventory around fulfillment center floors to inventory stowers and order pickers. In Amazon Unbound, Brad Stone estimates that the average pick rate in a fulfillment center jumped from 100 items an hour to 300–400 after the introduction of Kiva technology.

A Sober Look at Amazon’s Automation Drive

Because the Kivas are built like a hockey puck, they’re very simple. They move by wheels, they don’t need complicated systems to balance, and they can lift loads that are well beyond the capacity of any proposed humanoid robot. They are only capable of certain tasks, but they have a form factor that is optimized for what Amazon needs them to do.

9 Likes

My question got four answers so far, Amazon is the only investible company and they are not in the business of selling robots.

The Captain

Ah. Well, a quick google search reveals a number of publicly traded robotics companies, like Fanuc and ABB and Yaskawa. They’re just not U.S. companies. If you’re investigating how advancements in AI are being deployed in robotics, those might be companies to look at. For example:

A moment’s further research reveals that one of those companies (ABB) is probably off the list. SoftBank recently announced they were buying their entire robotics division as an AI move, making the same sort of move you’re talking about but on a much larger scale:

Anyway, as you can see discussed in that article, there’s tons of activity in the AI robotics space outside of the humanoid form factor. If, or as, AI makes robot “brains” more capable, there’s going to be demand for machines that can use those brains - and not only, or even especially, in a humanoid case.

2 Likes

No, I’m not.

I sure didn’t need Google to figure that out.

Elsewhere someone asked: “What non-optimal financial decisions have you made in your life?” to which I replied:

Investing in technology instead of investing in cash flow.

Don’t invest in EVs, invest in the EV maker that produces good cash flow.

To you I reply, Don’t invest in robots, invest in cash flow.

Which company can I invest in that can compete with Tesla in humanoid robots? Show me the money!

The Captain

What is Tesla’s 2025 cash flow from humanoid robots?

6 Likes

All of them. Tesla earns no cash flow from humanoid robots. They don’t yet have a working humanoid robot product. All of these companies that are generating real cash flow from non-humanoid robots are competing with Tesla, and since Tesla hasn’t created a product entry yet, they’re all generating vastly more money than Tesla is on “embodied AI.”

Nor is it certain that humanoid robots will ever^^ be a material product line, rather than ending up like the solar roof or even the Cybertruck - something that’s technically feasible and that Tesla nominally produces, but doesn’t generate enough volume of sales that it has any significant effect on Tesla’s business operations or cash flow.

^^ Subject to the caveat that even if humanoid robots end up being nothing but an immaterial niche product in any relevant time frame, Musk might still have one of his other companies buy a few hundred thousand at discount in order to nominally satisfy his compensation threshold.

5 Likes

Do you guys have any other world shattering news for?

The Captain

Look, the answer to your question doesn’t involve anything world-shatteringly new or surprising. Tesla’s trying to develop a robot to take advantage of advances in AI - so is every other company that currently makes robots, and there’s scores of companies that are newly trying to do it. If you don’t want to talk about the technology or those companies’ other attributes, but just comparing the money that they make selling these robots to the money generated by Tesla selling humanoid robots, you get a ridiculously simple answer: because Tesla isn’t making any money selling humanoid robots.

So maybe the way you framed the question - looking only at cash flow - isn’t one that will be useful at this stage in the development of the technology.

6 Likes

Check out Symbotic SYM.
It plans to be the AI that orchestrates robot warehouses.

Receive wholesale product in, high density storage, then distribute that inventory to end distributors.
The robot middleman automating the supply web.

The warehouse AI will control individual robots that perform specific task.
Sort of a gigawarehouse.
It fits within my definition of what is a robot.

Current stock price, about $65, is the same as 2 years ago.
Yahoo finance:
Market cap 37B
Total cash 1.2B
Debt 32M

SYM is making the rounds of the YT stock channels that YT feeds me.

I do NOT own SYM.
I’m just mentioning it. I do like the concept.
I might own some … In the future.

:alien_monster: :factory:
ralph

2 Likes

THANKS! Looks interesting at first glance:

The problem with the Fool numbers is that some are annual and others quarterly. $61.5 million investment vs. $1.79 billion is only 3% but if the $61.5 million is quarterly and the $1.79 billion annual then then it’s a steep 14%. Will have to look at the financial statements.

EDIT:

CBOE has the wrong company name, I just contacted them

Ticker SYM is now Symbotic Inc.
CBOE option chains shows SYMS CORP. (SYM), a defunct off-price retail clothing store.

The Captain

1 Like