The equation was always very simple: one fighter, one weapon: an archer for each arrow, a rifle for each sniper, a pilot for each fighter…, but the artificial intelligence is beginning to break that relationship.
The US Army has just tested new software capable of coordinating several drones simultaneously using a single human operator. The technology, based on artificial intelligence and designed to work even when communications are limited or being interfered with, brings closer a scenario that until a few years ago seemed like something out of science fiction: swarms of drones acting as a coordinated unit.
The idea is simple to describe, although extremely complex to implement. Instead of one person piloting each aircraft individually, the operator defines a mission and oversees the operation. From that moment on, Drones exchange information with each other, distribute tasks and adapt their movements in real time. It is a difference similar to that between conducting an orchestra and personally playing each instrument. The first military drones essentially functioned as drones. Each move depended on orders sent by an operator.
More recent systems use algorithms capable of making local decisions. If one of the devices detects a target, it can share the information with the rest of the group. If another drone is disabled, its companions they can automatically redistribute the mission. Even in situations where communications are degraded, drones can continue to execute part of their assigned tasks. Technology tested by the US Army and named Gremlin-Xis based precisely on this philosophy: reducing human workload and allowing a single person to supervise multiple aerial platforms at the same time.
Engineers usually compare these systems with the collective behaviors observed in natureto. No starling commands a flock of thousands of birds. No ant controls an anthill. No fish coordinates an entire school. However, surprisingly sophisticated collective behavior emerges from simple rules and communication between individuals.
Drone swarms pursue the same principle. Each device knows its position, shares information with the others and continually adapts its behavior to the environment. The result is a distributed network capable of exploring terrain, monitoring targets or coordinating attacks much more flexibly than a traditional centralized system.
Perhaps the best way to understand the importance of this advance is to remember how computers worked half a century ago. Previously, a single central machine did all the work. Today the Internet distributes information among millions of devices. “Military swarms” follow a similar evolution. Palladyne AI, the company responsible for this software, points out that in August HE will carry out “an exercise involving more than 9,000 people operating in contested and multi-domain environments.”
Although military applications receive the most attention, the technology could end up being used in civilian settings as well. Coordinated drone swarms could inspect infrastructure, search for survivors after an earthquake, monitor forest fires or map large areas of land in a fraction of the time currently needed. The difference is that, instead of controlling each device individually, one person could direct dozens of drones through a single interface. For now, neither the army nor Palladyne AI have specified the number of drones that each operator will be able to control, but there is talk of dozens of them.