Podcast: Teradyne Robotics group president asks what's the point of humanoid robots

Key takeaways

Humanoid robots are inefficient for factories; wheels and modular cobots offer more speed, reliability, and value.
Traditional robots fail in high-variability tasks; new AI-driven cobots adapt better to changing factory demands.
Labor shortages and aging populations drive the need for intelligent machines to sustain manufacturing output.
In factories, practicality outweighs appearance—modular, task-focused robots outperform humanoid designs.

In this episode of Great Question: A Manufacturing Podcast, Ujjwal Kumar, group president of Teradyne Robotics, says automated humanoids are cool, but they're not terribly practical. Why force a robot to stand on two legs when wheels could get them across a shop floor faster and with greater stability? Why create a standalone, human-like machine that needs batters when you could install an arm with a power source? In this formerly live conversation with IndustryWeek chief editor Robert Schoenberger, Kumar discusses why humans are so much better at some tasks and why a humanoid robot isn't a great solution.

Below is an excerpt from the podcast:

RS: Can you talk a little bit about some of the downsides of using the human form factor on the factory floor?

UK: Sure. So, when you look at it, most of the work which gets done on a factory floor, a big chunk, you don't need to move out of the cell for efficiency. Like, we as human beings, we move out for bio break and coffee break. If you have a robot working there, doing welding, palletizing or cutting, some of the CNC work or assembly work, you can pretty much, stay at one place with your six arms or move on a rail and manage most of the stuff.

And there are other things which just needs to be moved efficiently from point A to point B, the, limitations of a human form factor with so many degrees of freedom, which by design will need lot more battery power. Higher degrees of freedom, more joints also means, lower reliability, lower in quality, with no real value for the customer.

So, I feel that there are limitations for that, human form factor to the kind of stuff we won't get done in a factory.

RS: I think one of the I comments you made was that, you think about, wheels versus legs. If you have a flat factory floor, wouldn't you want wheels versus legs? Because they're more efficient. And I think of the maybe the genius of George Lucas. And, if you imagine that original C3P0 wandering around in Star Wars that is kind of what I see in a lot of these humanoid robots that I've seen a sort of slow, jerky movement. Especially at the Automate show, some of the robots could do a nice little dance, but if you say that's okay, go pick up that box, it was move over to the box. Stop slowly. Bend down. Slowly grab it. Slowly pick it up. Where a person could have done the same job in a 10th of the time, versus R2D2, which might have been the less humanoid of the two, but it was definitely more utilitarian. Could move around quickly on his wheels, easily outpace his counterpart. Yes, it's fiction, but it does seem to match what we've seen on factory floors.

UK: And I love these, the, graphics here. As you can see, the a machine which needs to work on a factory floor doesn't need the restriction of an AI on the top two hands. Exactly 20 inches apart. Legs, which are inherently unstable, like us human beings. Right? We are inherently unstable. As soon as you take the battery out of us will collapse. Like how humanoids do. Do we know, is that the most efficient way to do things in the world? In in a factory floor? I don't think so. We as humans, we did not evolve to be most efficient factory workers. We evolved for safety and other social needs.

So, as we think about what is the best way to move some of the repetitive, dirty, dangerous work over to machines? Human form factor doesn't look like the most efficient one to me. Having run factories for more than two decades. More practical than, say, people who are willing to buy things just because they look like humans or talk like humans. I want what I want to get work done at the same time.

RS: As we talk about some of these limitations, there have been companies many times over the years who've tried to automate certain functions that people do very well. I think one of the most famous examples, again, comes from Tesla, with the Model Three rollout, they tried to heavily automate their factory in California, especially that final trim area where people are bending past obstacles on cars, attaching different things. And it's high variability work, lots of finding tools, finding individual pieces. The robots turned out to be just very, very bad at it, and people could do it very kind of almost mindlessly. It was very simple to go from one, one process to the next up. The robots really struggled with this. So, you would think that if you could have a machine that replicates a human, there would be this, kind of one-to-one exchange. Is it just a technology difference that we can't get there at this point?

UK: No. I think what Tesla tried to do is a classic, case study, the robotic technology, which have been around for last, 50 years. Those are designed for high volume, low variability applications. Tesla tried to use that for their factory, where they had a lot of variability and frequent design changes. That is where you use more advanced robotics. So, the kind of, portfolio we have in, say, Universal Robots or Mobile Industrial Robots, these AMRs and cobots, they are designed to run AI and other software based applications to handle variability. And, these kind of the new generation of advanced robotics can be deployed and redeployed easily the way we human beings we get trained and retrained, easily.

So, I think the Tesla example was an older technology used in today's problem of variability and will, human beings, be better served to do that kind of work? Absolutely. My issue is, do we have enough humans to run our factories and keep our economies running. With the demographic changes, aging population, our biggest issue is if we want to maintain our quality of life, will need to figure out where do we bring in those more intelligent, machines to help do the kind of work which we human beings we were doing.

About the Podcast
Great Question: A Manufacturing Podcast offers news and information for the people who make, store and move things and those who manage and maintain the facilities where that work gets done. Manufacturers from chemical producers to automakers to machine shops can listen for critical insights into the technologies, economic conditions and best practices that can influence how to best run facilities to reach operational excellence.