WASHINGTON — The Defense Advanced Research Projects Agency and Dynetics proved they can catch an X-61A Gremlin drone in flight and bring it aboard a C-130A mothership. But the next step will be more difficult.
In a conference call with reporters Monday, Dynetics’ Gremlins program manager Tim Keeter said the program now must prove it can recover multiple drones repeatedly, reliably and quickly enough for real-world operations.
For DARPA, Keeter said, making the American program work at an “operationally relevant rate” would mean being able to recover four drones within a half-hour. He called last month’s successful test — the Gremlins’ fourth such deployment — a milestone for both the program and unmanned aviation, adding that the effort could “dramatically expand” the military’s ability to carry out distributed airborne operations.
He also expressed confidence that with the program over “the biggest hump” of proving airborne recovery can work, it will be able to reach its goal.
“Now we feel the momentum and the wind in our sails, and we’re ready to get this thing moving towards an initial operational capability,” Keeter said.
If the program works as intended, he added, it could one day allow the military to deploy “swarms” of drones into denied areas. That way, the mothership could stay out of harm’s way while the drones gather intelligence, conduct targeting or perform other functions with their sensors in dangerous airspace. The drones also could be deployed early in a conflict to overwhelm an adversary with numbers, he noted.
And eventually, Keeter said, the military could use artificial intelligence to allow these drones to behave autonomously, albeit with minimal supervision. However, the Gremlins program is not directly using AI right now for its flight demonstrations.
“What that might look like in a larger-scale conflict would be multiple sensors working together and fusing, that the Gremlins would carry into a contested environment, where right now even one operator could control up to four of those,” Keeter said.
If one Gremlin in the swarm is shot down, Keeter added, the rest of the drones could adjust for its loss and still carry out the mission.
Late last year, DARPA carried out a series of flight tests in which nine attempts were made to recover three Gremlins, none of which were successful due to greater-than-expected turbulence that kept the drones from locking into the docking station. After those tests, Keeter said, the program analyzed the data to identify ways to lessen the turbulence that stymied those recovery attempts.
Brandon Hiller, Dynetics’ chief engineer for the Gremlins program, said the program saw several key improvements put in place since the late 2020 test to make it possible to catch the drone in the docking bullet, shut down drone operations and reel it in. The finetuning was almost entirely in the program’s software, he said.
More work remains until the Gremlin program meets DARPA’s goal of four recoveries in 30 minutes, Hiller said — but the program thinks it has a path forward.
“We’re not there yet,” Hiller said. “But … we know what needs to change in the system, both from a hardware and a software perspective, to get us to that point.”
Keeter thinks the program will be able to carry out the quick recovery of multiple Gremlins during their next deployment, though he could not say when that might take place.
After that, the program will soon move onto its next phase, which will demonstrate how the drones would perform on a mission. That would mean “putting it together with payloads and single operators controlling multiple vehicles in a live-virtual-constructive environment, performing different aspects of the missions [to] showcase what this capability could really do and how it could help transform our military,” Keeter said.
Additionally, the Gremlins recovery system can be adapted to other aircraft, meaning the program won’t always need to use the C-130A as its mothership, Keeter said.
The first successful airborne Gremlin recovery took place Oct. 29 at Dugway Proving Ground in Utah. It was the third flight in that test. The first Gremlin that took flight that day had a power problem that developed quickly, and the drone was lost before its parachute recovery system could fully deploy. The cause of that Gremlin’s loss was first reported by Breaking Defense.
Keeter called the loss of that Gremlin “a great disappointment,” but one that was part of the learning process. The team saw that either an anomaly or equipment issue caused it to go wrong, and team members switched to a different configuration to avoid that problem in the other flights. He declined to go into further detail without DARPA’s approval.
“When you have a system with this many complexities and things interacting together, you come prepared for lots of options to fly with,” Keeter said. “Certain options may put you in areas where you’re more likely to encounter other issues.”
The second drone that flew in the test was not recovered in midair, instead parachuting to the ground.
The third drone was the one that was recovered and was turned around for a second flight within 24 working hours. However that drone’s second flight was not recovered in midair, and it too landed on the ground “fairly softly” with the help of its parachute, Keeter said.
Stephen Losey is the air warfare reporter for Defense News. He previously covered leadership and personnel issues at Air Force Times, and the Pentagon, special operations and air warfare at Military.com. He has traveled to the Middle East to cover U.S. Air Force operations.