Building autonomous robots 10X faster with 3D printing and a modular robotics platform

What if building autonomous robots was like picking options on a car? I’ll have the vision module, the speech module … better give me a wheeled transportation package for this one, a flying navigation module for that one … I’ll take LIDAR and plus a vision module … maybe a physical arm with a gripper and a object recognition library … and so on …

Ohmnilabs offers an modular robotics platform that lets companies configure autonomous robots. They have customers like Google, Amazon, Apple, and Toyota, and they 3D print components so you can test new versions quickly. The idea: pick your hardware, pick your software, build your robot.

In this episode of TechFirst with John Koetsier, I chat with Ohmnilabs’ CEO Thuc Vu about how it works, what’s possible, and how much faster/better/cheaper this model is …

Scroll down for full video, audio, a transcript, and to subscribe.

And … here’s the post on Forbes

Watch: Building robots with 3D printing

(Subscribe to my YouTube channel so you’ll get notified when I go live with future guests, or see the videos later.)

Subscribe to the audio podcast

 

Full transcript: Can you configure autonomous robots like building a shopping list?

(This transcript has been lightly edited for length and clarity.)

John Koetsier:  One robot is cool. A thousand robots though, is a revolution. It’s all very well to bring in a robot here or there, even everywhere … but to transform our economy, we need hundreds and hundreds of robots that are customized for different jobs — automated, deployed, coordinated. How are we going to do that?

OhmniLabs says it has the solution with Ohmni Modular Robotics Platform, which customers can use to design their own robots from “base to brain,” as OhmniLabs says, and then 3D print them in Silicon Valley. To learn more, we’re chatting with CEO, Thuc Vu.

Welcome, Thuc!

Thuc Vu: Hello, John. Very nice to be here. 

John Koetsier: Hey, super happy to have you. Excited about this, very cool stuff. Maybe let’s start here: what is the Ohmni Modular Robotics Platform? 

Thuc Vu, co-founder and CEO of OhmniLabs

Thuc Vu, co-founder and CEO of OhmniLabs

Thuc Vu: It’s a really unique and agile approach to robotics development. So we’re building out this platform with four pillars.

One pillar is a set of ready-to-use, off-the-shelf robots. And then the second pillar is a very rich, comprehensive library of robotics components, sort of like the building blocks that you can plug and play on the robots. And then we tie them all together with our unique 3D printing approach to manufacture these robots that allow us to operate almost like a software company.

And so now we can go to great extent to customize our robot on the hardware side, to view different functionalities depending on the needs of the verticals or the market. And then we also have the software side that, you know, sort of like a library that people can program different applications on top of it. And so, yeah, this is how we can really go to a different variation of robots in a much shorter time and cost. 

John Koetsier: So, that sounds really cool, because I might be a manufacturing company, maybe I’m a warehouse/logistics company, whatever. I might have some specific robotic needs but zero ability to actually make the robots that would help me in my business.

I can kind of come to you. I can use some web-based system to design it and you’ll make it — I assume the entire robot isn’t 3D printed, there’s probably certain specific components that are? 

Thuc Vu: Mm-hmm. Yeah, so we use a hybrid approach, right? Whatever the simple kind of off-the-shelf components that we can buy, we’ll use them. You know, why not? Because they have been designed well.

But there are intricate, complex pieces of the robot that actually 3D printing is a great application to do this, because we can design them with stronger, structural design in mind or, you know, different kind of parts that can — instead of injection molding, you have to do multiple parts and combine them — we can just bring up in one shot.

And so that’s really opened up a whole new playing field for us. 

John Koetsier: And for other manufacturing methods like injection molding and others, you generally want to do that for fairly large quantities, right? I mean, it’s expensive to set up. It’s cheap to run once set up, but your dies might be tens of thousands of dollars in some cases, correct? 

Thuc Vu: Yeah, absolutely. And you get locked in. You get locked in for, another 10,000 robots before you can do any changes to it, right?

But for us, we can manufacture at a volume of, say, tens for our customer. And so you just want, you know, you just need tens to test out a market and see how well it worked. You don’t need 10,000 in the beginning.

And we work with the customer; we’ll grow with them to develop this fleet of robots that really customly fit into what the customer needs, right? And so that’s the beauty of it. 

John Koetsier: I love that. That’s kind of a minimum viable product scenario there because, I mean, like you think you want to throw a robot at a problem; you think that might be a good solution, but you definitely don’t want to invest millions of dollars in inventing your own robots or hiring somebody to build your robots for you.

You want to be able to test it with $10,000, or $20,000 to $30,000, and I assume that your platform allows that?

Thuc Vu: Yeah, absolutely. We believe that one of the big roadblock for mass adoption is the cost versus the value. Right? And so, anyone who wants to deploy robots need to have enough value for the cost of robots. And so for value, devil’s in the details, right. Sometimes, you know, maybe it’s 10% that you need to change the robot to really make it work for your factory. And, you know, other people wouldn’t be able to provide that, but we can because of the way that we design and build robots.

And then on the costs side, we can, we just really try to make it affordable with all of this. And so that’s how we believe that we can roll this out.

John Koetsier: Let’s talk about what’s possible. Who do you ship robots to? And what kinds of robots do you ship? 

Thuc Vu: So it’s been really interesting, since the pandemic we have seen an explosion of new use cases. People are very creative in terms of how they want to deploy/apply the robots in their businesses.

We have seen our robots being used for virtual tourism, basically allowing people to travel again, without flying. It’s fascinating. So you can dial into a robot in Paris to go shopping and then, boom, after you’ve done shopping, you can dial into your robot in New York to check out the art gallery, and then go see your parents in Silicon Valley or Tokyo. So it’s fascinating what people can do.

Some of OhmniLab's robots

Some of OhmniLab’s robots

We have seen a lot of application in healthcare and education, especially for remote learning. You know, kids,  stuck at home, right, so now they can actually go to classes through the robot, spending time with their friends. Industrial use cases is actually picking up very nicely. People are using our robot to dive into a manufacturing floor to provide remote support, remote training, inspection … and now one interesting recent use case that we’ve been exploring with is UV cleaning.

So a team from Stanford and Target reached out to us, wanting us to build a robot for them that can use UV component to clean commonly high-touch surfaces. And we managed to do that just within one month from the time we talked to them to like we had something to demo.

You know, that’s incredible speed that we’re talking about, rather than years like the traditional approach, right? 

John Koetsier: Is that typical? Is a month from, ‘Hey, I need a robot for X’ to, ‘Wow, I have a robot for X,’ is that typically a month? And how does that timeframe compare to somebody who has to like build it a traditional way? 

Thuc Vu: Yeah, traditionally an approach will take two to three years [and] tens of millions to get to the market with a robot. And so with our approach, because we have all the building blocks, right, sort of, we just like combine them together with our printers, with our 3D printing technology.

And so it allows us to operate a lot faster, you know, we’re looking at three to six months and under a million to get to the market with a new type of robot. And so this is the speed that, you know, yeah, unheard of before. 

John Koetsier: Interesting. Interesting. What are the parameters like? What are the capabilities? On the hardware side, what are the big things that you can sort of piece together like Lego bricks to make something? And on the software side, what are the capabilities there? 

Thuc Vu: Yeah, we have over 200 modules spanning across hardware/software to cloud system.

And on the hardware side, typically you have all the major components of a robot, right: the drivetrain, the brain, the processing component, display unit, audio units, you know, those kinds of things.

And then, but they have variation, right? If you need a more beefy robot, we have a stronger processing component for you. But if not, we’ll stick with the weaker one so that we save you cost, for example. And then on the software side, we have teleoperation.

We have a really good teleoperations stack, basically allowing you to control and remotely control the robot or the machine. We have some big corporation that actually [is] experimenting our technology stack to control a remote excavator [laughing]. 

John Koetsier: Oh wow.

Thuc Vu: To drive an excavator with our module … it’s fascinating. I cannot go into too much details, but yeah, it’s interesting. And then autonomy, that’s a big part of our fleet, allowing the robots to basically navigate moving around a space, mapping it out, remembering the route that they need to take later. Going forward, we have some really interesting services in the pipeline.

For example, perception, for the robot to recognizing different objects. And an arm … that’s what we’re really excited about, allowing the robots to manipulate moving objects in the real world. 

John Koetsier: Okay. Very, very cool. So grippers and all kinds of different things like that, wheels for getting around; put all the pieces together, figure out what you need. How many variations of a robot can you design in your platform? 

Thuc Vu: So we have launched about 10 different robots so far, within the span of two years and a half.

John Koetsier: But if you threw everything into a blender with all the different components you have… how many can you create?

Thuc Vu: It’s limitless. So, you know, you can … yeah, it’s just fascinating.

And when we need to change certain model of our robot, we just need to send it the CAD files to the printers … and boom, the next day we have a new variation that we can experiment with, right? And so it’s just, yeah, the incredible speed. 

John Koetsier: Wow. So, do you focus on a couple of verticals? You mentioned quite a few telepresence use cases, and so there’s a lot of verticals that telepresence could work in. But what are prioritizing over the next little while for development, in terms of capabilities and verticals you’re targeting?

Thuc Vu: Yeah. So, I think UV cleaning is another vertical that we’re getting a lot of interest and demand. And so that’s one area that we’ll be focusing on next.

And then, you know, the autonomy and the perception of robots really unlocking many different type of use cases. So, indoor delivery is actually really interesting, bringing a parcel from the lobby to the right office. You know, you have a hundred floor- or seventy-floor building, it’d take a lot of effort to go up and down just to bring those parcels. And so that’s a very well-suited task for a robot.

And then security is also another area, an indoor security robot and it just roam around, identify abnormal activities or people who shouldn’t be there. And then the last one is sort of like the concierge/receptionist use case, especially for hospitals and office buildings. Like, you know, U.S. has been doing very well with the vaccine rollout, but other countries, they’re still struggling with the new surge of COVID. And a lot of times when the patient comes in at the hospital, the receptionists don’t know whether they should put them in like a special lane or just let them breeze through, right? And so a robot will actually [be] reducing contamination risk at that, yeah.

John Koetsier: Okay. Wow. So, speculate a little bit for me, you know, I don’t know, five, ten years out … I come to OhmniLabs and can I design pretty much anything I can think of? You know, I want a wheeled robot. Actually, I want a flying drone; I need this level of perception ability; I need autonomy in 3D space, or I don’t; I need the ability to synthesize speech or at least to speak to somebody; and I need a gripper or something like that. I can load this drone up and design whatever I want — is that the future you envision? 

Thuc Vu: Yes, absolutely. I think five to ten years is a bit short [laughter] to spend across drones and— [crosstalk]

John Koetsier: C’mon, work harder. [laughter]

Thuc Vu: Yeah, and indoor robots, right?

But what we wanted to really do is to democratize robotics technology. Right? Basically open it up so that many people can work together; can build different types of robot. Sort of like driving the ubiquity of robots in this world and bringing value right to the user, rather than just us building robots. We want to be an ecosystem that can use our modular platform, and you can build your own flavor of robots for your own customers. 

John Koetsier: Okay.

Thuc Vu: Why not? Right? That’s how we can really drive the innovation, the new types of robots in this world to bring value to everyone.

John Koetsier: Interesting. So it’s kind of, it’s not Intel inside, it’s OhmniLabs inside and you design and build whatever you want. You deploy your fleet; there you go. Anybody can come to you, not just end users, but also people who want to customize a robot and then resell it. 

Thuc Vu: Absolutely. Yeah. We’ll be the engine powering your robot company.

John Koetsier: Excellent. Excellent. Well, Thuc, thank you so much for your time! Very cool, what you’re working on, and look forward to learning more as you build it. 

Thuc Vu: Yeah. Thank you for having me. 

It’s official. If you make it down here, you have to subscribe

Made it all the way down here? Who are you?!? 🙂

The TechFirst with John Koetsier podcast is about tech that is changing the world, and innovators who are shaping the future. Guests include former Apple CEO John Scully. The head of Facebook gaming. Amazon’s head of robotics. GitHub’s CTO. Twitter’s chief information security officer, and much more. Scientists inventing smart contact lenses. Startup entrepreneurs. Google executives. Former Microsoft CTO Nathan Myhrvold. And much, much more.

Subscribe on your podcast platform of choice:

 


Want weekly updates? Of course you do …