A new AI tool from GDIC that quickly fuses data from multiple air-defense sensors could transform how militaries defend against emerging aerial threats such as hypersonic missiles and drone swarms.
The proliferation of drones and highly maneuverable missiles has made conventional air defense targeting difficult and increased the complexity of threats from the air. But advances in technology have also resulted in a wider range of data—from radar, satellites, or drones—to inform and help defenders better spot incoming threats. The challenge is to collect all of that data, even in places where adversaries will be trying to jam communications, quickly analyze and integrate it, and send it back to give the operators on the ground enough time to take action. GDIC tested their Defense Operations Grid-Mesh Accelerator, or DOGMA, tool against that problem at the military’s TREX event in August.
DOGMA ingests available data about what’s coming from the network of sensors and looks across all the available communication paths to pick and choose the best path to transfer the data along the difficult “first mile” near the front lines.
Brandon Bean, GDIC’s director of AI/machine learning (defense division), told Defense One, “It could be anything from man-portable radios all the way up to [low-earth orbit or geosynchronous orbit] satellite connectivity.” The tool can quickly send data to a Combined Air Operations Center and then back again, applying AI to the different air threats the sensors are picking up.
Past the first mile, DOGMA can route the traffic to the cloud using commercial pathways.
“If you’re in INDOPACOM, the first available [Amazon Web Services cloud] availability zone would be Tokyo, Singapore, Sydney. Once you hit that availability zone, they then route you on a dedicated path using the undersea fiber cables that they have. So you’re not comingled with other commercial Internet traffic,” Bean said.
To test how well it works, the company set up a geofence around an aircraft hangar at Indiana’s Camp Atterbury during the August exercise. “The objective here was to replicate a [Taiwan’s air defense identification zone] incursion. So that geofence was about 1,500 feet tall and about 500 meters wide,” Bean explained. “In the end, what they found is that they could provide a high probability estimate of where even a highly maneuverable threat is headed with about 30 seconds of lead time to the operator.”
Getting that information across that “first mile,” where the adversary is trying to block all communications paths, is one of the key innovations, Bean said.
“We didn’t have an [electronic warfare] threat in the simulated environment. But what we did was… periodically, without notice, would just start pulling plugs on all of our communications paths.” That resulted in a communication disruption of just 33 milliseconds, as the system found the next-best communication path, essentially unnoticeable to the operator.
New missile technology, especially highly-maneuverable supersonic and hypersonic missiles, are hugely difficult to defend against. But Beam said the system could be applied to help operators defend even against these missiles. However, he said, there’s a big challenge to overcome—namely “the amount of training data we had available to us. We put in a request with the FAA national office to get about three months worth of data from their major regional FAA centers. And that request is still in queue.”
The post New AI tool for air defense takes on advanced missiles and drone swarms appeared first on Defense One.
