EXECUTIVE INTERVIEW: Metamaterials for military radar, invisibility cloaks, and more
StoryFebruary 13, 2024
Engineers at Echodyne have enhanced the capability of small radars for military applications such a counter-UAS [uncrewed aircraft system]. In a podcast I did with Tom Driscoll, co-founder and CTO of Echodyne, he and I discussed how his team’s metamaterials electronically scanned array (MESA) radar system does things that traditional active electronically scanned array (AESA) radar systems cannot, the company’s unique business model, and how it leverages commercial off-the-shelf (COTS) components like Xilinx FPGAs [field-programmable gate arrays]. Tom also shared details about his background in metamaterials and how he once designed an RF invisibility cloak using them. Edited excerpts follow.
McHALE: Can you please give us a brief description of your responsibility within Echodyne?
DRISCOLL: I’m one of the two co-founders of Echodyne and I’ve served as a CTO since our foundation in 2014. I was also the managing director of the technology incubator we came out of and as such was one of the handful of people that invented our core technology. I’ve worn a number of hats over the eight years of growth and development of Echodyne. I started off pretty much changing light bulbs and picking furniture like everybody does in a startup company. I’ve been chief architect of almost all of our products and [ran all of engineering] until we divided research and engineering. Today, I still run all of our research and development team, looking at what’s next in some advanced capabilities.
McHALE: Many people I talk to in industry – from primes to COTS suppliers – have told me to go check out your company and what you guys are doing. Echodyne CEO and co-founder Evan Frankenberg told me at DSEI in London last fall that Echodyne’s radar product came about while you were looking at an alternative to expensive AESA radars and landed on using metamaterials to design your MESA radar. I also saw that you were a visiting scientist at the Duke University Center of Metamaterials and Integrated Photonics. What is it about metamaterials that makes for effective radar systems? How did Echodyne come up with the MESA radar?
DRISCOLL: I probably can’t say this for certain, but I think it might be true, that I’m probably the first person in the world to get a Ph.D. in this new physics field of metamaterials. And that was just serendipity. I started my Ph.D. at UC San Diego and looked at what all the research groups are doing. And I really liked this group with a professor named David Smith that was talking about electromagnetic metamaterials, and new ways to manipulate light and radio waves and microwaves. That’s really where I jumped in. Following my graduation from UCSD, I got courted to come out to Duke University and help do research there. Duke had really made a name for itself as the center of the academic metamaterials world. We did some fun stuff out there.
I wasn’t an author on this paper, but I think one of the most famous experiments [was] an invisibility cloak. Now, it’s an invisibility cloak for microwaves, but [it] might be good for stealth fighters or something like that. It certainly got a lot of attention. And then after Duke, I got courted to come out to Seattle. There was a company here that was trying to make a name for itself as the center of the industrial and commercialization metamaterials world. That’s where we got this tech incubator started. I’ve been in metamaterials for probably going on close to 20 years now [and] we’ve identified a lot of cool things we can do with it.
When I came out to Seattle, the charter was to move away from academic interests and figure out how we could make money with metamaterials, either licensing it or through startups. We pretty quickly identified antennas as this really rich area for innovation. Antennas are in everything. It might not sound amazing, but if you can make an antenna and a cellphone just a couple percent better, that’s worth a lot of money because there’s so many cellphones. So, we explored a number of different antenna-related areas. I really became enamored with radar and radar antennas. It’s this rich, deep technical field that you can learn in for your entire life and still not know at all.
We asked the question, could we use metamaterials to improve radar? And the answer was, yes, we could use metamaterials to create an antenna that electrically steers a beam in the same way as that phased-array active ESA does for what we thought was orders of magnitude reduction in the cost, size, weight, and power (SWaP) of the net system. We’ve been maturing and improving [the technology] for the eight years since.
McHALE: Regarding RF invisibility cloaks, you mentioned fighter jets as an application for them. How would that work? Would the metamaterials deflect the signals used to detect the fighter, effectively making it invisible? Am I oversimplifying?
DRISCOLL: Well, we never did it for a fighter. To the best of my knowledge, no one’s done it for a fighter, I think there’s still a significant technical and engineering gap that needs to be closed before anybody could. But if somebody had done it for a fighter, I probably wouldn’t know about it.
The difference between true invisibility and stealth is kind of subtle. If something is invisible, it has to work from all angles, you have to be able to turn it around in 360 degrees in both directions. And it has to still be invisible. That’s different than camouflage, because I could paint the background on one side of an object and it would look like it blends in. But when I turned it, I would see that in the turn off it was only painted on one side. It’s different than stealth because stealth works most of the time by trying to absorb the energy and avoid reflecting it back so that you don’t reflect an energy back, then it’s difficult to detect you. Invisibility catches the energy, bends it around the object, like water flowing around the rock in a river, and then sends it out the other side completely unperturbed. You never even knew it was there. There was never a ripple.
McHALE: So, it’s kind of like in science fiction, when spaceships cloak, they are bending the light, so to speak?
DRISCOLL: That’s exactly right. That’s what we’re doing with these nanomaterial devices, is we’re bending light. Light sometimes at the wavelength of optical light, but more often at the wavelength of microwaves and radio waves. That’s because they have longer wavelengths. It’s easier to manipulate structures at those longer wavelengths.
McHALE: Are there tradeoffs between the MESA and AESA solutions?
DRISCOLL: There are absolutely tradeoffs. Engineering is just the art of tradeoffs. And most of the time an invention in engineering is just another tool in the toolbox. And sometimes it’s the right tool, and sometimes it’s not.
The output, the beam that’s formed by our metamaterials ESA, and a traditional phased-array, or AESA are very similar to each other. It’s the way in which we form that beam that’s different. In a traditional phased array, you have these elements called phase shifters, hence the name. And you need to base those phase shifters at one-half of the wavelength away from each other. You do that typically in some second grid, and that forms an array. At every one of those sites in that array, the phase shifter picks the right phase, you adjust via some voltage or some digital signal. Then collectively, all the sites in that array form a narrow beam in a specific direction. That’s how you electronically scan the array.
In the metamaterials approach, we actually have a much denser array, instead of spacing them at half of a wavelength, we spaced them at about a tenth of a wavelength, so they’re five times denser. Then instead of having any phase shifters, what we do is we just turn each cell on or off. That binary collection of ons and offs creates what’s technically called a diffraction grating or an optical hologram. It’s an RF optical hologram, but it’s still a hologram. The beam that comes out is still pretty much the same as the beam that comes out of that phased array. But we don’t need these phase shifters. It’s simple in the way a CD is simple, digitally, compared to an analog tape recorder.
McHALE: So your secret sauce, that nobody else has come up with, is how you turn those things on and off?
DRISCOLL: That’s right. It’s the way in which you design that dense array. It’s the map that helps you understand which of them you turn on and off. And then a different pattern of ons and offs steers a beam in a different direction. So, by changing that pattern of ons and offs, that’s how I electronically steer my radar beam around a wide field of view or field of regard.
McHALE: How long has Echodyne been in business?
DRISCOLL: We started the company in December of 2014. A true statement about the radar market is that it’s incredibly fragmented. The radar for a car that keeps you from running into somebody is wildly different from the radar that’s on a fighter jet. And that’s wildly different from the radar you would put on a fishing boat. There are all these submarket verticals and when we started the company, we said, we’re not actually sure at the outset which of these markets we’re gonna go tackle. We spent a little [time] just doing R&D, improving our key performances, and then also doing market assessments.
We released the evaluation kit after about 18 months of business and shipped that evaluation kit to a bunch of people in all of these different submarket verticals. We let them play with it. That gained us some credit, some reputation that yes, in fact, this technique worked. It also allowed us to start good business-development conversations and figure out what their needs were and how many could we sell to that need.
The first markets we identified to productize on were drones, and this was launched as a dual product. It was launched as a product for drones that are in the air to enable them to use one of our radars to fly safely. This is what’s called a collision-avoidance radar. It scans the airspace for that airborne drone. It tells you if there’s a Cessna or some other small drone that you might run into and alert to so you cannot run into it simultaneously. We sold this in a ground-based version that could be used to protect against drones – whether it’s a just careless usage, like somebody’s flying for fun near an airport, or a nefarious usage, like somebody’s flying drugs into a prison using a drone. That was our first product. We launched that in 2017.
McHALE: By dual use do you mean in terms of import-export controls? Or do you have a different meaning?
DRISCOLL: I actually mean both of them: I only said one, but both of them are true. I meant dual use in that we call this airborne product EchoFlight for drones that are flying in the air. Then we’ve got this ground-based product we call EchoGuard that’s for a different set of customers to protect against drones from the ground. But it is true that Echodyne builds dual-use products in the defense commercial sense. We build and ship commercial products, but one of our largest customers tends to be the military and the DoD [Department of Defense].
McHALE: When I was speaking with [Echodyne co-founder] Evan at DSEI, he referenced that your radar was on a counter-UAS system on display at Anduril’s booth. Are counter-drone solutions a typical defense application Echodyne?
DRISCOLL: We do a lot of counter-drone, counter-UAS. I would say it is our single-largest market vertical.
McHALE: He also told me that your radar is a COTS radar, available off the shelf. Do you use a lot of COTS components as well such as single-board computers, FPGAs, etc.?
DRISCOLL: Almost everything in our radars is COTS electronics. And we in turn become COTS because [the radar is available off the shelf] and shipped within a week of the purchase order. The only thing that’s not COTS is the design of our metamaterial antennas. Obviously, there’s some custom sauce there. But the materials it’s built out of are standard such as circuit-board materials and then we put down Infineon and SkyWorks parts on top of those circuit-board materials. So, there’s no material science in metamaterials. Maybe it’s not the best word we could have named this as a field. Everybody always thinks that it’s in a cleanroom doing thin films, but it’s just a circuit board to us.
We do love the Xilinx [SoCs, systems on chip], which we build into SoMs [systems on module]. We buy those with the Ultra Scale Plus Zynq chips on it from a supplier. They sell that SoM to us as COTS because there’s a deep art to the design of high-speed digital stuff. We can save a little bit of time, money, and effort by just buying that from an external vendor.
Our circuit cards, the baseband, power-processing circuit cards are full of conventional parts that you’d get from Digikey. Our RF circuit cards are pretty much the same as well, just from a different set of vendors.
McHALE: A popular topic in our industry is AI [artificial intelligence]. Sometimes I think AI is just a marketing ploy, but in many design processes it’s not. Do you use AI or ML [machine learning] in your design process at all?
DRISCOLL: Well, for a long time, I kind of stood by a saying that AI was just the latest impressive thing that any computer can do. There was a point in time where I think people were saying that spellcheck was AI, even though it’s quite algorithmic. But I may have to back away from that snarky saying because I do think that the new transformer attention-based nets are something kind of interestingly new. It’s a definite generational leap forward. It’s not just the latest thing computers can do, it’s something new.
We don’t use any AI in our design process. It’s an interesting question if we could, but we don’t. And we do use less AI and more ML in some of our signal processing. But no neural net black boxes today. We are doing some interesting new research looking at using transformers and attention nets for some parts for signal processing. But it’s research, so I won’t say very much about that.
McHALE: We cover the modular open systems approach (MOSA) mandate and the Sensor Open Systems Architecture (SOSA) Technical Standard quite a bit. Echodyne is a member of the SOSA Consortium. Are you seeing more requirements related to MOSA and SOSA?
DRISCOLL: We’ve joined this consortium because it’s part of our ecosystem. It’s the responsible thing to do and these are people we talk to a lot. Our interface is not [SOSA aligned]. We looked at all the options out there [and determined] to get the most out of our radar, because it’s sort of new and sort of different, we had to write a custom API, although it is an open API. We’re not looking to lock up ecosystems. We encourage everybody to roll their own integration to it. But in terms of what we’re hearing, yeah, MOSA is pretty much the buzzword of the day with DoD program managers and program office. So, some form of it is definitely flowing down. We’ll pretty quickly figure out how to bridge the gap and compatibility between our API and most of the formats.
McHALE: When Echodyne was founded you said you explored at many different markets for your product. Who is your customer today? Is it mostly DoD? Do you sell directly to primes or to the end user? Do you compete with the primes?
DRISCOLL: I don’t think we compete with the primes. That’s hard to imagine. We’ve sold radars for applications ranging [from], in dynamic range and breath [beyond the DoD for applications such as] United States southern border security. For that Anduril tower that you mentioned, we are the radar that unlocks that tower and enables it to know where to point its camera. When we shipped them serial number one of EchoGuard, they put it on one of their towers, and they said, it was like somebody turned the lights on. During that first week of using [our] EchoGuard radar, they caught like 10 times the number of illegal crossings as they previously had.
Our uses range all the way from that, which is larger-scale, to somebody who’s using one of our radars to measure the speed and velocity of outbound golf balls. We’ve sold systems for a lot of different applications.
I can lump our ambitions really into three segments: the defense military market, the federal civilian market – like work with Customs and Border Patrol, and then critical infrastructure. Critical infrastructure [covers] everything from prisons to airports to stadiums that need systems to detect and to counter drones. [For these areas solutions scale] all the way down to what’s called perimeter intrusion detection systems, or PIDS. If you follow [U.S.] news, there’s been an increase in vandalism of electrical-transfer stations. Now, every electrical-transfer station of a certain size or above a certain criticality or above has to have one of these PIDS to detect if somebody has hopped the fence and is going to go do something not nice inside that electrical station. Those three verticals really capture 95% to 98% of our business, the DoD, [plus] sensitive and critical infrastructure verticals.
McHALE: Your MESA radar is available off the shelf. But on your website, you have a section for custom design: How much of what you do is custom?
DRISCOLL: Good question. Let me answer that by also returning and answering part of your previous one, which I skipped over. I said, I don’t think we’re a competitor in any way with the primes. I stand by that. I mean, any program of record is going to be owned and run through a prime. That’s just because they’re the trusted partners that have owned this ecosystem for decades. So, if we’re on a program of record it’s as a component supplier as a subcontractor to and through a prime and that’s where we want to be. We want to sell OEM radar sensors, and almost nobody can use an OEM radar sensor by itself. There needs to be a system around it, needs to be software integration, there needs to be display, there probably going to be other sensors, there might need to be effectors that do something about a drone or a target when you see it. Building out of that system, and interfacing with systems of systems is the work of the prime integrators. We’ve worked with just about all of them, we’re friendly with just about all of them. I know all of them have our product. We love it when we can sell the radars we have designed and built to them and that meets their needs.
But oftentimes, these primes pursue pretty sophisticated systems that can evolve radars that are beyond our design, understanding, or even just our resource capacity. But they still recognize that MESA technology is something new, and something disruptive. We get approached quite often with the question: Would we design a MESA antenna that they could use as a subsystem and plug it into their radar. They build all the radar back end, they do all the processing, they do everything, except we apply that phased-array substitution, that MESA ESA.
When we do this, it’s typically around an opportunity with a sizable scale to it. We’re not a design shop. We don’t really want to be coin-operated to just do these designs for an NRE basis, and then leave them be. We’ll partner with prime integrators, or really anybody who’s got a big opportunity, to design a MESA antenna that we can sell to them at scale with the goal being product revenue around that subsystem.
McHALE: What capabilities or do you have on your R&D roadmap? Are you going to bring back the invisibility cloak?
DRISCOLL: Yeah, I don’t know who I can sell an invisibility cloak to; I’m not sure who’s gonna pay for it.
One of the cool things about this type of electronic steering array is it really unlocks a lot of software capability. If you think about the ability to instantaneously point this beam in any one direction: Now you’ve got this control loop, that software inside the radar can utilize that to measure things better, measure things differently. We’ve got a lot of software work ahead of us, adding new features, improving performance, operating better in challenging environments, like really dense cities, where there’s a bunch of cars and potentially power wires everywhere. Or at sea: Ocean clutter is an old and well-known adversary for radar. Sea waves are very challenging to deal with. So [we have] a lot of software work. Right now we’re doing the market research and diligence in each of these three verticals to see where we’re going to push our product portfolio. We certainly see an opportunity to move up in weight class in the defense and DoD space and build a longer-range, higher-power radar.
We also think that to unlock some of the truly high-scale critical infrastructure opportunity, we might want to build something that’s even a little smaller and cheaper than our lowest-end radar.
