On a recent Wednesday morning, Ph.D. student Ryan Farell was climbing to the top floor of Darrell K Royal-Texas Memorial Stadium at The University of Texas at Austin — he was on a mission to detect drones that were attempting to remain undetected while simulating attack routes.
The air sorties were part of a U.S. Army exercise called DiDEX 3 (defense in depth experiment), which ran Dec. 6-10 in the northeast corner of downtown Austin. The exercise tested the most advanced technologies for spotting and stopping drone assaults in urban areas.
“This is an exciting and critical event for national security that allows our students and faculty to better understand how the Army operates in real-world scenarios,” says Seth Wilk, director of UT Austin’s Defense Research Advancement office.
The event was sponsored by the Combat Capabilities Development Command C5ISR Center, a part of the Army Futures Command (AFC). The command, which chose Austin for its headquarters in 2018, partners the military with private businesses and academic researchers to speed innovation and modernization of military technologies.
Farell had climbed to the top of the stadium because he was given the opportunity to be embedded with soldiers using sensors and control software from 10 different vendors. He, Maruthi Akella, a professor in the Aerospace Engineering and Engineering Mechanics Department, and four ROTC students evaluated how rapidly and effectively each technology detected the drones, as well as how user-friendly each was for soldiers to set up and operate.
Farell, who is part of UT’s Oden Institute for Computational Engineering and Sciences and an expert in deep reinforcement learning and data science, is exploring the use of machine learning to combine diverse streams of sensor data — such as video, infrared, radio waves and lidar — into a single operating picture. “It might be that it’s very hard to detect something over the infrared spectrum but that you can hear the radio frequency, because it’s getting blasted through a building,” he says. “Or it might be that you can’t detect the object with radar, because the drone has been modified to evade radar, but you can detect it visually. The question is how to integrate all of them.”
Akella, an expert on autonomous robotics, designs systems to help unmanned space vehicles and mobile robots communicate with one another and collectively respond to threats, independent of human direction. At DiDEX 3, he cross-checked the data from a variety of sensors to see how reliably they helped identify potentially hostile vehicles. “One system detected a drone, but I want an independent confirmation,” he says. “I will seek out a different data stream to make sure that, yes, there was a vehicle at that point at that time.”
Both researchers say DiDEX 3 will help them create technologies by giving them a close-up of today’s capabilities.
“Because a lot of the research is being done by private companies, it’s not obvious what the current state of the art is,” Farell says. “Being able to participate in an event like this, you see what the outputs of their systems are, which aren’t otherwise accessible to you.”
DiDEX 3 was the third in a series of trials in countering unmanned aircraft systems (UAS). Others were held in Virginia and New Orleans. Over the course of the week, mini-swarms of up to seven drones were launched in more than 400 simulated sorties from 40 locations. Scenarios included strikes on sporting events at the stadium and an inauguration at the Capitol.
A key attraction of Austin was its concentrated downtown, says Bill Newmeyer, chief of the Sensor Assessment & Interoperability Branch of the C5ISR Center.
“When you get into dense urban environments, the sensors that are supposed to detect these drones are limited by tall buildings,” he explains. Because each type of sensor has limitations, software layers a variety of them into a single picture of what’s happening in the airspace — a strategy called defense in depth.
From sensors at such spots as the stadium and the Texas State Library (next to the Capitol), a team of engineers based at the Sheraton Austin Hotel, on 11th Street near I-35, processed raw data into a common operating picture. They employed artificial intelligence to identify what kind of vehicle each UAS might be, whether it was hostile, and what track it was headed on. They streamed the resulting picture to soldiers on the ground, who could view it on smartphones attached to their body armor.
An added attraction for AFC was the ready cooperation of the city, state and university. “The government, academia and industry partnership in Austin has been a huge success,” says John Klopfenstein, team lead for the Unique Mission Cell, a C5ISR Center group that conducts experiments to discover and reduce system vulnerabilities.
Wilk agrees. “It demonstrates how academia and the U.S. Army can successfully work closely together on modernization priorities. The participation and engagement of our students and faculty enabled immediate improvements, which the Army integrated into their systems throughout the week.”
For Farell and Akella, the next step will be to submit detailed reports to be shared with both military personnel and the vendors whose technologies they tested that week. Their assessments will help contractors develop the next generation of drone defenses while also helping the AFC design future technology experiments like DiDEX 3.
At the same time, they’ll be looking out for new research possibilities: practical problems that UT’s expertise might help to solve. “Where are the technology gaps?” Akella asks. “Where are the knowledge gaps that would motivate further foundational research we can do at the university?”
Farell looks forward to being part of that research. At UT Austin, he appreciates the chance to help design the future of America’s defenses, even while he’s pursuing his degree.
“It’s a great opportunity to stay relevant,” he says. “We get to see the current technology and its limitations, and we get to suggest future research initiatives. I’m excited about that, because I want to be involved in research for a long time.”
