Autonomous collaboration: Getting robots, drones, self-driving vehicles, and smart machines to work together

Autonomous collaboration

We know how to get people to work together. How do you enable autonomous collaboration … or getting robots, drones, autonomous vehicles, and smart systems to collaborate?

In this edition of TechFirst with John Koetsier we chat with Kumardev Chatterjee, the founder and CEO of autonomy-as-a-service startup Unmanned.Life. Unmanned Life has already deployed autonomous drones with the city of Vienna, enabling the fire department to automate some parts of search and rescue operations. The company is also working on projects with Walmart, Swiss Post, Teléfonica, and Deutsche Telekom.

Some of the topics we cover:

  • autonomy as a service
  • software that enables hardware
  • Tesla
  • autonomous collaboration in a machine-first world
  • machine communication protocols
  • robot-machine cooperation
  • and much more …

Listen: Autonomous collaboration

Better yet, subscribe to TechFirst with John Koetsier wherever podcasts are published:

Watch: Autonomous collaboration

Full transcript: Autonomous collaboration

John Koetsier: We know how to get people to work together. How do you get robots, drones, autonomous vehicles, and other smart systems to collaborate? 

Welcome to TechFirst with John Koetsier.

No jobs are simple. Take ordering a pizza and getting it: you’ve got to phone somebody, or use an app, or be online, that order has to go somewhere, somebody has to take it. Some bread needs to be prepared, some dough, some toppings need to be put on. That gets finished, goes in an oven, goes to a delivery person, he picks it up/she picks it up, drives it to your home. All that is coordinated, but everything takes steps and processes.

If we’re going to get to an automated world, an autonomous life if you want to say it that way, how do you get smart machines in the kitchen, in delivery, in production, all those things working together? How do you enable that? To chat about this we are going to bring in our guest, who is Kumardev Chatterjee. He’s the founder and CEO of Unmanned.Life.

Welcome Kumardev!

Kumardev Chatterjee

Kumardev Chatterjee

Kumardev Chatterjee: Thank you, John, welcome everyone. Very nice to be here and speak about these topics which are very close to my heart. 

John Koetsier: Wonderful. Talk to me, as a bit of an intro a little bit, how you see the future of machines working together.

 Kumardev Chatterjee:  I believe we are entering what I call an ‘era of autonomy’ or an ‘era of the autonomous economy.’ And what that effectively means is that today we have a humans-first in most jobs and most sort of processes, whether that’s industrial, or retail, or even food delivery you just mentioned that, medical delivery … and my vision is that we are going to move to a world where machines-first. So the first thing you think about when you’re thinking about delivery of food would be, not who’s going to come and knock on your door, but which drone is it going to be today, right? And the first thing you’ll think about when you’re like, ‘Oh, I’ve got to go to a hospital, or I’ve got to get a checkup done,’ is basically which machine room with which machines are going to be looking at you. 

And today, this might sound a bit overtly futuristic perhaps, like “Back to the Future” type of futuristic, but the reality is that this is happening, as you know and as I know, across the world at an increasing pace. And why is that? It’s because machines are not only the future in terms of productivity and efficiency, but today we need them. Look at the Covid world and it’s obvious that we need to have less and less human interaction for things where humans don’t need to be put at risk. And that’s where machines are perfectly suited for that. 

John Koetsier: And we have huge challenges with that, right? Because you see a lot of machines being created, you see a lot of robots being made, delivery robots, automated machinery being put into factories, other places like that. But often we’re thinking of a very discreet job or a very specific task for the machine, and not completing an entire task. So we have something that might help us plant a seed but maybe not also harvest, or it might help us drive a nail but not build the whole house, correct?

Kumardev Chatterjee: Absolutely, John, spot-on. And let’s ask ourselves the question from a very standard standpoint. Why is that? Because if you go to a greenhouse, you’ll see that the person who’s pulling the seed out of a plant is not the same person who’s watering the plant, and it’s not the same person who’s actually sort of working on that seed and turning that into something else. It’s a team of different human beings performing different tasks. They might be skilled at different tasks, but in a particular process they do one particular task in general. Now in our world, we all do multiple tasks sometimes at the same time, but for most industrial, or retail, or normal processes, one person does one particular job. And if you were to replace that with machines, exactly, it’s not enough to have one machine or a couple of machines that do different things, you need to have a team of machines, robots, drones, whatever you call them, working together just like a human team would do. And so we’ve got to get over that to be able to enable the scenario where an entire process run by human beings can be now run by robots and perhaps some human beings. 

John Koetsier: And to do that, we don’t just need the hardware. We need some communication protocols, we need some interoperability protocols, correct?

 Kumardev Chatterjee:  Music to my ears, right. So the history of this goes back quite a while. A lot of people have tried to build the ‘everything machine’ right? And you know that as a tech journalist, as somebody in the field, and they have failed. And for good reasons, is that hardware is not only hard to do, hardware is fundamentally from the physics standpoint, built to do certain things. And when you think about a team of hardware, a team of robots working together, number one is that point you mentioned: communication. 

How do you communicate? I mean, just take a rugby team or what you guys call football, right? It’s rugby the rest of the world. So take a rugby team or a football team, and you’re like, ‘Let’s communicate and let’s figure out how to do a scrum.’ But to communicate somebody has to take leadership, and sometimes leadership, the baton has to be passed on. These are questionable things for human beings to perhaps do, but very hard for machines to do, particularly just hardware. So you need some sort of brain that’s not hardware, that’s software that is able to manage those communications and bring it together.

A very simple question I’d ask everybody, ‘If there’s five people in the room and you didn’t know each other, and you didn’t speak each other’s language at all, and, but everybody, I mean, we were told at the same time in your language, “Get out of the door at full speed, one at a time.” What are the chances you’re going to do that without colliding? Zero, right? 

John Koetsier: Yeah.

Kumardev Chatterjee: And so that’s exactly what happens if you have a hardware-based… you’ve got to have a hardware solution. You need to have a supervisory brain, a software that’s able to actually talk to each of those robots and these machines, we don’t understand each other and don’t share a common language, but not only, they need to be directed in a way by someone, or something that is able to see what needs to be done and what each one is individually doing. And this needs to be done at very fast, very fast and adaptably, so if something starts off and then the position for different robots change, someone needs to understand that and readapt what the next step will be. That can only be done by software mathematically from a computer science perspective. And you know, we can be a scientist, we know that can be done only through software. 

And so that’s the missing bit, John, which is why a lot of these ‘let’s-do-everything- machines” haven’t gone anywhere. I mean, I know that you are a great Tesla fan and Tesla undoubtedly is at the edge of the automotive world when it comes to EVs and autonomous driving. And if you look under the hood of a Tesla, of course it’s impressive hardware, but there’s super impressive software. Ahead of the game on all the software issues, right? And that is the key. You can only have fantastic hardware performance if, and only if, your software is ahead of the curve and is able to make the hardware think.

John Koetsier: That’s a really, really great insight. And it’s funny because when you started talking about Tesla where I thought you were going for a second, was Tesla was going to build the machine that was going to make the machine, they were going to build the totally automated factory. They pulled back from that because they lacked the ability to actually do that. It is something that I assume will happen in the mid to near-term future, but they couldn’t do that.

But you’re absolutely right, great hardware is hobbled with poor software.

I’m probably going to offend a bunch of people here, but I’d call out Samsung as one of the companies in the world that need to learn that and probably have learned that.  What I wanted to ask you is you’ve actually, so you’re not just building something out of thin air here, you’re actually implementing things. Talk about what you’ve built so far and what you’ve implemented so far.

Kumardev Chatterjee: So the idea of autonomy of a service is no longer a concept, it’s something that’s been deployed by major governments around the world. The city of Vienna is deploying the world’s first fully autonomous drone service for search and rescue. So our drones will take off from fire engines and will go and search for people who are lost or who are in distress, and then first responders will take that data and respond to that.

And now with Covid, we’re now deploying autonomous disinfection solutions. We just won the award today, we’re one of the winners of the EUvsVirus Hackathon. I published that on my Facebook. And what we’re saying is that you can have autonomous disinfection for hospitals and large public places, and even indoors, because our drones can do outdoors and indoors, just like a human team would do, but there are no humans, so they’re not at risk from Covid or either to themselves or to others.

And these drones can be deployed whenever you want. 

John Koetsier: Yes.

Kumardev Chatterjee: And that’s because of the software platform that we’ve tested and deployed at Walmart, that we are testing to deploy with Walmart, with Swiss Post, with Teléfonica, Deutsche Telekom and other places, and the city of Vienna among other things. So, definitely the future of autonomy is software and not hardware, but it’s more than that. It’s about software that is hardware agnostic.

John Koetsier: Yes.

Kumardev Chatterjee:  And that layer of software can just run robots without you having to think about what robots and how. And we are a part of the Walmart accelerator and effectively what we do there is you’ve got robots working together in a team, in a warehouse and moving parcels at full speed. So I could show you a video, but I think might not fit in the format of the live interview. I’ll send that to you later on. And you’ll see these robots just like human beings moving around in a warehouse. There are no tags on the floor, there are no motion capture cameras, there’s no GPS obviously, it’s indoors, right?

And these robots have to work like human beings, so they’ll have to figure out where to go and pick up a parcel, they’ll have to wait for the parcel, politely take the parcel, and only once it has been loaded start to move at a very high speed, up to five meters per second, get to the other end of the warehouse to a point that they didn’t know what that point was when they picked up that parcel, because you know, they didn’t know where they had to go, and somebody would be directing them through this whole process.

Looks like an orchestrated dance, humans and robots working together, what I call kind of like a ‘next generation Swan Lake’ effectively. 

John Koetsier: Interesting. It’s a super interesting concept because you can imagine how an Amazon would love to do something like this. Amazon, of course, bought the Kiva robots, the company that built them, and they’re probably one of the larger installations globally of workplace robots. And frankly, from what I hear from people who work in Amazon distribution centers and the repetitive stress injuries that they have, it would be a good thing.

Kumardev Chatterjee: No, absolutely. And I think, it’s interesting that you mentioned Amazon, a lot of people when we started off said the same thing, ‘You know, look guys, what about Kiva? What about Amazon?’ And I said, ‘What’s interesting about the Amazon approach doing that, and they tried that too with Kiva, is that you have to build a warehouse that’s Kiva compliant.’

John Koetsier: Yes.

Kumardev Chatterjee:  And the rest of the world doesn’t want to do that. And Amazon obviously for its own competitiveness doesn’t want to give you their technology. But we as a software, we can just say any company, whether that’s Walmart, or any of the other large American companies that some of them which you are working with, just takes our software and starts deploying that in their warehouses and it looks cool.

John Koetsier: I love the idea of making it work for factories that are not built for automation because we’re going to be working with people for large portions of time. You’re going to have some robots that can do some parts of it, but you’re not going to be able to invest in robots that do everything right away. You probably don’t have the money to be able to do that, but being able to have a system that enables people to work with robots makes a ton of sense.

You mentioned drones. What are the drones doing? 

Kumardev Chatterjee: So the drones are working together to respond to an emergency. So they are flying and they’re giving data to first responders so that they can figure out what’s going on.  

John Koetsier: Perhaps it’s a firefighting situation or something like that, you can send the drones, get some imagery and some data on what’s coming or what’s happening, and be able to respond to it quicker. 

Kumardev Chatterjee: Exactly. Yeah, that’s exactly it. So when we started out there’s basically two assumptions that had to be taken on head-on. So the first one we just talked about, can hardware do everything, can you build a hardware-everything machine? And the answer to that obviously was that no, you can’t, right? You just cannot build a hardware-everything machine. The other assumption or the other challenge was, can you do a software that’s able to talk to different types of robots? So this is not just about talking to two types of drones from two different organizations, or two types of drones from two different manufacturers, it’s about, can you control a drone and a robot together with a robotic arm and can you compose a fully autonomous team where there are different pieces of machine working together?

John Koetsier: Interesting.

Kumardev Chatterjee: Right? So that is something that we are working on and we’ve done a bunch of that, we have deployed robots and drones to work together. 

John Koetsier: Is there a standard for robot-to-robot communication? Does it all go to the cloud first and then come back down? Does it go device-to-device? How does that work?

Kumardev Chatterjee: So there’s both types of operations. So you’ve got companies that do like clear path and some others that build robots and they have proprietary protocol where different robots can work together. If you can get this done and we’ve got it done, and then scale that up, suddenly that autonomous economy where you have robots or teams of robots … (audio lost) … sometimes with humans, sometimes without humans, becomes a reality, right? 

John Koetsier: Interesting. 

Kumardev Chatterjee: And so from the very beginning, the first challenge was, yeah, so very first challenge, how do you solve the hardware issue? Different types of hardware from different manufacturers that don’t talk to each other, no standards, whatever, how do you crack that? Number two was how do you do different machines working together? And I often had this analogy, if you were like Superman, we could go up and then come down and walk. Well, that’s just fantasy. Can you do that in reality? Right, and that’s what we’ve done. And where are we going from here is actually trying to scale that up and do a large number of robots working together in teams. 

John Koetsier: Well and the challenge is as well that the human body, I mean, it’s kind of like a Superman to a robot. I mean, the number of different things we can do, we can sit, we can stand, we can walk, we can run, we can handle with our fingers, you know, versus most robots are built to do kind of one or two or three of those things maybe, at the very most, and combining those in especially a mobile platform with power is really, really challenging.

Well, Kumardev I think we’re going to have to call it for now. We’ve had some internet challenges and other things like that, but saw some cool things. Really, really do appreciate it. 

Kumardev Chatterjee: No, John, thank you very much. I hope that I was able to share some insights and of course with a better internet we could have done a bit more. But yeah, I mean, I think what’s interesting is that we’ve managed to now enable something that was simply not possible before.

John Koetsier:  Thank you again, Kumardev, it’s been super interesting. And for everybody else, thank you for joining us on TechFirst. My name is John Koetsier, appreciate you being along for the ride. Whatever platform you’re on, like it, subscribe, share, comment, all of the above. If you’re on the podcast later on, please rate it and review it, that’d be a massive help. Until next time, this is John Koetsier with TechFirst.