A hawk soars over a field, its wings tracing an intricate dance. Each flap, each tilt is not random but a combination of simple motions that can be broken down into components. Collective behavior in the air may seem intuitive, but behind it lies strict mechanics.
Scientists recorded the hawks' movements using cameras that track the position of every point on the wing and tail. They then applied a method that extracts the main repeating patterns from these recordings — a kind of "alphabet" of flight. It turned out that just three such letters suffice to describe wingbeats, and a fourth is added for turns. All birds share the same set of letters, though each has its own handwriting.
This approach allows predicting how a bird will move next, even if it does something new. It's like being able to complete the missing part of a pattern once you know its basis. The model does not rely on assumptions about the physics of flight; it is built directly from data, so it reflects the bird's actual behavior more accurately.
In the end, we get a simple and clear picture: a hawk's flight is not infinite variety but a few basic elements that combine. This perspective not only helps us understand how birds fly but could also be useful for creating flying robots that move naturally.
Describing a hawk's wingbeat takes just three numbers, and a turn adds a fourth. That's comparable to how a few notes form a melody.
In the Strugatsky brothers' story "The Predatory Things of Our Time," there's a device that reads human movements; here we have almost the same thing but for birds: we break flight down into simple elements.