Drones Learn to Fly in Formation Without GPS — by Acting Chaotic

Most drone shows and delivery fleets depend on a steady stream of directions from a ground station or GPS satellites. Lose that signal — inside a building, in a disaster zone, or in a war zone where signals are jammed — and the whole formation can fall apart. This paper asks a simple question: what if the swarm carries the directions with it?

Here's how it works. One drone is designated the 'embedded leader.' It stores the starting point and the destination, works out the route on its own, and sets the pace using something called a Chua oscillator — a math formula famous for producing wild, never-quite-repeating wiggles. Crucially, the timing stays inside a safe speed range, so the arrival time can still be predicted by the swarm itself. That one speed number gets broadcast to a few other leaders, and everyone else just watches the drones nearby, holding position the way a spring and a shock absorber hold a car's wheel steady.

The math says the leaders' errors shrink steadily and the followers stay close, so the whole group ends up parked inside the destination zone — as long as the final formation fits. Ten computer simulations with 30 drones confirmed the arrival times landed within the predicted window. The fun part: the chaotic pacing made the swarm's arrival much harder for an outside predictor to guess, compared with drones flying at a constant speed. The trade-off was slightly sloppier spacing between drones — but still within safe limits.

The honest catch: this is pure simulation. The researchers assumed perfect communication between drones, and the paper doesn't tackle what happens when a link breaks or a drone drops out. No physical drones flew. So this is groundwork, not a product — but it points toward swarms that keep working when the signal doesn't.