Matternet’s Package Delivery Drones
Amazon’s talk of package delivery drones may just be pie-in-the-sky, but start-up Matternet has already begun testing a delivery-drone network in developing countries
Tekla Perry: Hi, this is Tekla Perry for IEEE Spectrum’s “Techwise Conversations.”
Amazon’s announcement of Prime Air came out of the blue earlier this month, a lot like the service’s proposed eight-propeller drones would do. The company says that these drones, each the size of a large shoebox, could be delivering packages within a couple of years, if it can get clearance to use the drones in U.S. airspace. The reaction, to put it mildly, has been skeptical. And, indeed, drones swooping into urban environments, weaving around buildings and power lines, seems fraught with problems.
But Amazon’s idea isn’t exactly new. An organization called Matternet has been developing a network of drones. It plans to start deploying these in remote areas that don’t have reliable transportation systems. It describes its network as a “physical Internet,” and its drones would deliver important lightweight packages containing, perhaps, medications to take to patients in remote areas, and blood tests to bring back from these patients to the labs. Matternet has started testing prototypes. My guest today, Andreas Raptopoulos, is the CEO of Matternet and led the team that came up with the idea back in 2011. Andreas, welcome to the podcast.
Andreas Raptopoulos: Thank you very much for having me.
Tekla Perry: Why don’t you start by explaining what Matternet means by a “network of drones.”
Andreas Raptopoulos: The basic idea is to rethink our transportation infrastructure and try to ask the basic question: For countries that haven’t yet developed adequate road infrastructure, would it make sense for them to go through the whole cycle of investing in road infrastructure—the billions of dollars that would require and the many years it would take to build it—or is there a better technology to start resolving some of these transportation needs? And we looked at the space and what is happening in technology today, and we found this really exciting trend in the small UAV space. The octocopters, the quadcopters, the small fixed-wing vehicles that are all autonomous. So we thought, Would it be possible to create a network, a node network, where these vehicles could be transporting autonomously small goods between ground stations, and in that way set up a new type of transportation infrastructure, create a new paradigm of transportation that doesn’t rely on roads? This type of application, of transportation mode, would not only have applications in places that don’t yet have roads, like the developing world, but in many places here, in the developed world, like our cities and megacities, where we do have transportation infrastructure but it is very inefficient because of congestion.
Tekla Perry: So you envision yours could work in cities as well?
Andreas Raptopoulos: Yes, eventually. We figured it would make a lot of sense to start at places in the developing world, where there is a very high level of need and people are willing to take much higher risk in trying something out because they have a very, very pressing problem. And if you are able to solve it, you may be able to save a lot of lives or have a lot of positive impact in the region. And then over time, as we learn how to operate this, we can take it to rural places in the U.S. and, eventually, cities.
What Amazon is proposing, I think, is on the right track. The timing of it is unknown, but in any way that you see it, we think it is really necessary for corporations to start thinking, companies to start operating networks, to really understand how such a network can operate at scale.
Tekla Perry: You said it all got started in 2011. Can you tell me how the whole idea came about and how you got started on this project?
Andreas Raptopoulos: We started at a place in Silicon Valley, at the [NASA Research Park] called Singularity University. And it is a place where people—around 80 people from 35-plus countries—every summer gather there to talk about advanced technologies and how we can use them to address humanity’s grand challenges. And we saw that people who are locked in the cycle of poverty, of extreme poverty, which is about 1.4 billion people in the world, two-thirds of them do not have access to reliable transportation. So we understood that in order to really think of a world that is able to get these people participating in the economy, we really need to figure out a way for them to get access to commercial hubs and develop transportation networks around them.
So that’s how the whole thing started. The original vision was to start at a small scale and eventually go up in the payload capacity of those vehicles, carrying heavier and heavier goods, eventually coming to the point where we could do transportation of maybe groceries and other heavy products. What we found out, though, is that the world in some ways is already miniaturized. It is counterintuitive, but there are a lot of things in the small-payload scale. You heard the stats from Amazon that 86 percent of their packages are below 5 pounds? We found this pattern in many, many different places. So even with a very small payload capacity, you can really do a lot of transportation, have a lot of impact, in a place that doesn’t have any means of providing transportation.
Tekla Perry: Back to your network of drones, how far apart are each of the base stations?
Andreas Raptopoulos: This is still a design parameter for us. We have basically constraints of costs of the vehicle we are using, reliability. We want it to be quite reliable, and again, we want it to be able to transport certain payloads within a certain range. Initially, we put a stake in the ground of 2 kilograms over 10 kilometers, which would then place a requirement for those ground stations to be only 10 km apart. Then we started pushing the range further without increasing payload capacity. So now we are at the point where we can do 2 kg over 20 km, and we see a few good ways to increase that even further. The end game for the design of the vehicle is to allow us to not have ground stations when we don’t need it for delivery or picking up a load.
Tekla Perry: And these ground stations are necessary for recharging?
Andreas Raptopoulos: The ground stations, they come in different flavors, as it were. The fullest specification one is able to do battery and load swap automatically. The second version of the ground station is the one that is only able to do battery swap, and you need those stations in places where you don’t have a customer facing a need for that type of place. So these are more like the backbone of the system, and these are only able to do battery swap automatically. And then we have the third and the simplest version, of something like a pad that the vehicle identifies and lands on, without the need to do a battery or a load swap. So in the Amazon type of scenario, that type of pad would be the one that will be used to place in someone’s backyard, at some point in the future, and receive a package.
Tekla Perry: Can you describe the drones you are using now?
Andreas Raptopoulos: Yes. We are experimenting with different platforms. The one we used in the field last year is an octocopter; that means it has eight copters. The biggest costs about [US] $3000. Most of the vehicle components are off the shelf. We are optimizing the weight of what we are using in the package to extend the battery range, the flight range. And also, we are starting to think a little bit about the fail-safes we need to have on board, the redundancies we need to have on board. This is really where our work starts. These types of vehicles, these types of small drones we are using, they do not require piloting. They use only GPS and other sensors on board to navigate a certain path, and because we have fixed locations we fly in the network, we only fly between those marked locations, we either have a ground station or a pad, we are able to preauthorize the actual routes these vehicles fly. So we only do repeatable transportation, repeatable routes, on specific airpaths.
Tekla Perry: You started testing. Why don’t you tell me about those tests, and I’m also curious how long you think it’s going to be before you have something operating commercially somewhere.
Andreas Raptopoulos: We’ve done our first field trials last year in the Dominican Republic and Haiti, consistently with our vision of taking this first to places of higher need and places that really have a much higher urgency to try something like this. We went to those two countries, and we tried operations both in rural areas and in urban areas, in the capital of the Dominican Republic, Santo Domingo, and also the capital of Haiti, Port-au-Prince. We wanted to test different hypotheses of how the system would work, how people would react to it, and it was a very, very successful pilot; it beat our expectations hands down on most counts. And it gave us a lot of confidence to move forward and really try to think about a longer-term trial of the system, which we are trying to make happen now. And we are focusing on two different places in the world: in the Dominican Republic and in a small country in South Africa called Lesotho. The application in both cases is transportation of lab samples from remote clinics to hospital labs. The topology, the geography, and the different conditions around the trials are very different in the two places, but what we are transporting is the same in both cases. If we are successful in these first steps, then the logical next step is to really start using this type of treatment for commercial transportation.
Tekla Perry: So you are a for-profit company that thinks you can make a business out of this?
Andreas Raptopoulos: We started with the mind-set that we wanted to be as big as mobile telephony. Who would have imagined 30 years ago, when we had very cumbersome, very big mobile phones, only available to very rich people, who would have imagined that this type of telephony network would be available to everybody and it would be the biggest tool for economic growth in places like Kenya or Haiti? Our vision is really to do the same thing. The key difference that we have here is that when things fail in our field, there is a lot of damage that may be caused. Our view is that this is really going to be a new paradigm, and we are going to see all sorts of uses for it. Some of the uses are predicted already. The one that Amazon is talking about is one of the very straightforward ones. But then there’s going to be other uses that we can’t imagine today.
Tekla Perry: Did you think you’d have a competitor like Amazon quite so soon?
Andreas Raptopoulos: Is Amazon a competitor?
Tekla Perry: I don’t know, is it?
Andreas Raptopoulos: [laughs] We don’t know if they plan to create their own technology. Our plan as a company, our business model, is not to run networks ourselves. Our plan is to develop the technology for other people to set up transportation networks. So we see Amazon as a potential client rather than a potential competitor. Beyond that point, even if they are a competitor, our purpose here is to create a new industry. That has been our purpose from day one, when we started this. And in order to create a new industry, you need big players, strong players, that have an intense need around the problem we are trying to solve, to get into that space and start thinking about it, with real applications. So it is only good that Amazon is getting in.
Tekla Perry: Okay, then. Thank you.
Andreas Raptopoulos: Thank you. It’s a very exciting space, so stay tuned. I think there are going to be a lot of developments over the next few months.
Tekla Perry: We’ve been speaking with Andreas Raptopoulos from Matternet about a plan to develop drone networks to deliver important packages around the world.
For IEEE Spectrum’s “Techwise Conversations,” I’m Tekla Perry.
This interview was recorded 5 December 2013.
Audio engineer: Francesco Ferorelli
NOTE: Transcripts are created for the convenience of our readers and listeners and may not perfectly match their associated interviews and narratives. The authoritative record of IEEE Spectrum’s audio programming is the audio version.