Behind effective drone warfare, a living globe of geospatial data

Terry Gerton Derek and Will, thank you both so much for joining me today. It’s great to have you on the floor at AUSA. There’s a lot of attention here during the convention on autonomous weaponry, autonomous sensing, autonomous systems. When a military force might be in a battlefield, there’s a lot of different ways to get data. It could be a person, a soldier, it could be drone, it could a ground vehicle. But they are all going to report this data differently. Why is it so difficult to get a common picture?

Will Cocos So one of the big issues in warfare today is the reliance on drones and the full motion video feeds coming from those drones. What I think a lot of people don’t realize is the pointing accuracy of these drones is — it’s quite inaccurate. And you’re not able to extract a coordinate that’s useful for targeting; in some cases, even mission-planning. And so all of these different systems, multiple different types of drones and even multiple different types of satellite imagery, they’re all coming together in a battlefield commanders’ battle management system. And they are getting multiple different locations. And so the key thing we found is the need for essentially a universal, accurate datum that you can reference all of these inaccurate systems against. And that then allows you to take fundamentally inaccurate data sources, geo-register them, and then use them for things like targeting or mission-planning or whatever the use case is.

Terry Gerton So, Derek, when that happens in an operation, what kind of impact does it have?

Derek Stevens Things lined up, all those things become more efficient, they become faster. It’s really about reducing that error and increasing the trust. So not only for the soldier, but also for the commander to make decisions. But you know when things go wrong and things remain inaccurate, you could take the wrong path, you can target the wrong thing, and you know, it could be safety for a soldier as well. So things can go wrong quickly and accuracy and precision are incredibly important.

Terry Gerton How do we solve that problem today?

Derek Stevens I think it’s solved from individual systems. So we do what’s good enough for that individual system. I think you really start to see that problem when it goes across systems, or when it goes across forces, or it goes with international partners. But we’ve also had a lot of high-end equipment, especially as it relates to drones, where GPS, laser-guided munitions — these are things that have been trusted and validated and verified in many different exercises and events. So we’ve built up that confidence.

Terry Gerton And how much error can the battlefield tolerate before it starts affecting outcomes?

Derek Stevens It’s a great question. I think it just really depends on the echelon as well as whether it’s maneuver or fires. Small errors can have small mistakes, but they can also have very large mistakes. So it’s really just about, I think, understanding that error and the commander taking that into account for their risk assessment.

Will Cocos  I would say that the standard for Category 1 targeting, which is the most precise targeting, is six meters. So that’s the standard Vantor ascribes to. We are going to be far more accurate than that.

Terry Gerton It sounds like this is a huge math problem. All of these different surveillance feeds, all from different points, two different points, and you’ve got to pull them all together. How do you do that math in real time?

Will Cocos There’s a math problem for sure. I would actually say the biggest issue is a data problem. To be able to have the math to operate and determine a coordinate, you have to have a baseline reference system. And you can’t do that overnight. You have to collect multiple images from many different angles over many years of the entire globe. And then you have to knit those into a very accurate 3D base map. You have to have a very accurate pointing system in your satellite. You have ground control points. So all of those things have to come together over many years to map the globe. After that, yes, there’s math. The math is the relatively easy part. Not for me, but for the folks that build our technology, yeah. That’s the relatively easy part.

Terry Gerton So what happens has to happen with all of this data for the operator that Will was describing to actually trust the data that they’re getting?

Will Cocos Today, ultimately it has to be validated in an operational environment. Today, you know, that role is filled, in part, by NGA. They validate if satellite imagery accurate or not. How accurate is it? Can you use it in a targeting workflow? How would it be used in that targeting workflow? So, I mean, that’s the key thing. I think these processes have been around a long time, and they are made for a certain type of targeting workflow. You have to have multiple different images, those images have to go through a certain system, and then I use photogrammatic processes to determine an accurate coordinate. That’s slow. What we’re finding in today’s warfare — we’ve seen it with Ukraine, we’ve see it with other operational units we are working with, U.S. operational units — that targeting cycle is much faster. So, you might get a single image and then you need to go prosecute that target immediately following receiving that image. Maybe you receive it in less than 15 minutes. You don’t have the opportunity to take multiple images. And so you have to rapidly geo-register that and then feed that coordinate to a coordinate-seeking munition.

Terry Gerton When you’re working across services, different targeting systems, different processes, how do you make sure in the joint environment that they’re sharing the same information?

Will Cocos That’s an issue, you know, you see it with battle management systems. Different units are using battle management systems, they’re using different maps. And those maps, are they using different imagery? Are they using different feeds off of drones that all have a difference underlying reference system and a different underlying accuracy? The only solution that we’ve seen work is you have to have a global reference system. We work with another country, a partner of the U.S., figured this out a few years ago. And they were getting 3D feeds with full motion video, they were building 3D models, but it wasn’t accurate enough. And so to synthesize all of those, they leveraged our 3D backbone to build their map of the globe. And then from there, they’re able to leverage all the different use cases that they’re using it for. Viewsheds for specific weapon systems, targeting applications and so on.

Derek Stevens I think that’s a great point in that our data is also commercially — it’s shareable. It’s not hidden behind, you know, other systems or difficult to share. So I think that’s another benefit of all the things we’re talking about. It’s one way to do it.

Terry Gerton So Derek, let’s follow on with that for a minute, because Will was talking about the speed at which a unit needs to process targeting information. You’re talking about data in the clear where everybody can see it, but modern warfare moves really fast. So when you’ve got folks all the way from the tactical to the strategic level and things moving and sensors moving all the time, how do you get, in real time, that common perspective of the battlefield, the geo-registering? How does that all sync in real time on a modern battlefield?

Derek Stevens I think that’s the benefit of the scale of our data and our ability to update that not only with our own satellites, but our software allows it to run at the edge and use that reference datum. So you can have a reference datum on a handheld device out in the field. And if you’re collecting a drone, we can register an update to that reference datum. And we can obviously do that at scale in an enterprise or more central location as well.

Terry Gerton How sensitive is that collection of data and transmission of data to the network? What if the network goes down? How do you ensure communications and connectivity?

Derek Stevens We can run it at the edge on the hardware, so we don’t need to have connections. As long as they have the data that they want to align, whether that’s from the drone because they physically have it, or they have data on a USB drive, we could align it.

Terry Gerton It’s just so hard to imagine if you haven’t been there on a battlefield, all the things that are moving at once. And so, Will, everything we’ve seen in recent conflicts places a real premium on tempo and coordination. Where have you seen this ability to synchronize all of these feeds have a real impact or make a meaningful difference in planning or execution?

Will Cocos I think Ukraine is a great example. They derive targets from satellite imagery with multiple types and levels of accuracy, and they derive targets from drones, ISR drones with full motion video. Again, all of those things are inaccurate. They are running through a rapid target cycle between identifying the target and then launching a drone to prosecute it. So I mean, I think that’s the best example. They have to conserve their munitions. They have a limited amount. And so ensuring that they’re accurate, being able to geo-register that full motion video to a datum, a highly accurate datum that can be used for Category 1 targeting is critical in allowing them to move fast. So they are obviously, they’re doing this every day. They’re ahead of everyone on this. And frankly, this is not a solved problem. Yes, the technology exists. Just, you know, to be clear, this is not a solved problem for the U.S. Army or the U.S. military today. It is an issue right now. And I think as the U.S. military increasingly adopts drones and drone warfare and is using that, relaying coordinates from ISR drones to attack drones, that problem is only going to manifest further.

Terry Gerton Well, of course, we’ve just heard, you know, Secretary Hegseth announced the Autonomous Warfare Command. So, clearly drones or autonomous systems are going to be a key piece of everything going forward. What are the portions of the problem that still have yet to be solved?

Will Cocos The U.S. military does not have a global, highly accurate geo-reference system. That’s the simple problem. But just in terms of autonomous warfare, I mean, this is a military problem. This is also a commercial problem. Any autonomous system needs some highly accurate, 3D base map for training purposes if you want to train a drone to be autonomous. You can do that in a simulated 3D environment. That’s the most efficient way to do that. You have to have a highly accurate environment to do that. If you’re having a drone navigate, they need some base map, some base layer in 3D that they’re using to make their decisions. And then the third piece is autonomy. So if you have multiple drones, drone swarms operating in an environment, they have to a 3D base map to say, hey, I want to move from this location to that location and that’s good because, you know, this is the characteristics of that 3D terrain. And that has to be highly accurate to reality.

Terry Gerton And you just talked about drone swarms. Is it an exponential problem when you have so many different sensors trying to coordinate?

Will Cocos We don’t solve that swarming problem, but in terms of extracting accurate reference data, I think that can be. It’s a relatively, again, low compute, you need some data, you can export that data, but it’s relatively low compute. It’s a problem that can solved on the edge or solved on drones with relatively low compute capability.

Terry Gerton And Derek, we touched a little bit earlier on joint operations, multinational especially. And certainly here at AUSA, we’ve got all kinds of allied countries on the floor. Data is sort of multilingual, I guess, or it doesn’t matter which language you’re speaking. But how complicated is it to share this information across national systems, not just across our own internal systems?

Derek Stevens I think for that purpose, we use open standards for a lot of our data products. We use software to maximize the value of that data for those different workflows. We’re API-driven, so we can integrate with other systems or they can do the integration. But I think that’s what allows, from a warfighting perspective, people understand the value of the where. So understanding that it’s accurate, it’s precise, and how it’s being referenced is universal and that it makes their other maybe more unique or exquisite capabilities even more valuable.

Terry Gerton So when you think ahead, what is the future of this technology? Where does it go?

Derek Stevens  I think it’s enabling, obviously, the warfighting. It becomes a living globe, if you will, as far as sensors updating it, as well as that updated globe feeding back into those sensors. It becomes a cycle. I think from a drone perspective, as they mature and need to do additional things, understanding the environment that they’re flying into or how they’re impacted, all of that becomes just additional capabilities and additional layers.