By Stephen Losey
WASHINGTON — California-based defense technology firm Shield AI on Monday launched a new drone swarming capability called V-Bat Teams — one it hopes the Defense Department might use for programs such as its Replicator Initiative.
V-Bat Teams, which grew out of Shield AI’s experiments with the Air Force’s AFWERX innovation unit that culminated in a demonstration this summer, has at its core the company’s artificial intelligence pilot software dubbed Hivemind. These teams consisting of a handful of V-Bat aircraft are intended to operate autonomously in high-threat environments, without needing instructions or guidance from GPS or communications.
In an interview Tuesday at the Association of the U.S. Army’s annual conference in Washington, Shield AI’s co-founder and president said V-Bat Teams could be a “great fit” with the Pentagon’s Replicator program, which aims to field thousands of autonomous, attritable drones in the next two years to counter China.
Brandon Tseng added that V-Bat Teams would operate with minimal instruction from human operators, beyond the point where the humans tell them what target or mission to pursue. However, the V-Bats would notify humans when they notice something that needs to be brought to their attention.
Tseng said Shield AI conducted the first flight of V-Bat Teams in April and conducted a demonstration as part of the Air Force’s AFWERX autonomy effort in June. That demo, which was announced in August, showed how the Hivemind technology could launch and autonomously control a trio of V-Bats to monitor and surveil simulated wildfires.
V-Bat Teams now include four of the UAVs, Tseng said, but the company hopes to double that capacity every year — to eight in 2024, then to 16, and so on.
Hivemind’s autonomous software is, on its own, already able to control many more V-Bats, Tseng said. The limiting factor on V-Bat Teams comes in the operational logistics of launching multiple drones at once and then landing, he explained.
Sending large swarms of drones into the air isn’t a new trick, Tseng noted, as such demonstrations are commonly done at festivals or other celebratory events. But those are “brittle, dumb drones” that would fall out of the sky or automatically land if they were jammed, he added. Shield AI first focused on creating an intelligent, secure AI pilot in the form of Hivemind, and then the firm worked to add more drones into the mix to carry out different mission sets.
V-Bat Teams will first focus on maritime domain awareness missions, Tseng said, but their use could expand to include the suppression of enemy air defenses, strike operations, escort missions and logistics operations.
With their relatively low cost — Tseng said the V-Bat’s price point comes around the mid-six-figure range — V-Bats could even be used as decoys to lure out enemy fire and take hits that would otherwise target crewed aircraft.
V-Bats are attritable, Tseng noted, which means they could be sent into combat in considerable numbers and the military could easily weather their loss, while still being able to operate intelligently.
Tseng said some of Ukraine’s experiences have shown that throwing waves of basic drones at an enemy isn’t always enough to make a dent. “You need intelligent, affordable mass,” he noted. “Mass for the sake of mass is not helpful; it has very low returns.”
Shield AI aims to sell V-Bat Teams to all U.S. military services as well as foreign customers.
“You’re going to open up a new paradigm of these operations when you have this many aircraft that are able to do it autonomously,” Tseng said.
For now, however, Tseng doesn’t see V-Bat Teams fitting in with the Air Force’s plan to create a fleet of collaborative combat aircraft — autonomous drones that fly alongside crewed fighters such as F-35s. Shield AI is open to looking further into the collaborative combat aircraft program, he said, but what the Air Force has in mind differs from V-Bat’s design.
“We’re open to it,” Tseng explained. “But they want AI-piloted jets, and that’s what they’re focused on before they start thinking about how those AI-piloted jets work with other machines like a V-Bat.”
Read more here.