A Single Drone Control System: From Search to Target Capture ✎ express

Original: "Autonomous Tracking and Terminal Guidance of Moving Targets for Fixed-Wing UAVs"
· Wei-Hao Liou, Teng-Hu Cheng
arXiv:2607.12801 · 2026-07-14 · arXiv non-exclusive · independent analysis · 1 min
A single algorithm, like a falcon's instinct, guides a drone with a pivoting camera from search to precise target capture.

📄 Showing the "Simple" version — "Popular" is not ready yet. Add it to favorites to help prioritize it.

To make the drone act like a predator, engineers wove three hunting phases into one: search, track, and strike. No more task-switching: its unified logic runs seamlessly.

At its heart is a mathematical filter designed for guiding Apollo spacecraft. It predicts the target's path, smoothing out jerks and interference, like a falcon that already knows where its prey will bolt.

The camera, sharp as an eagle's eye, scans the scene, finding the target by contrast and brightness (just as a hawk spots a mouse in tall grass). To keep the frame unobstructed, the algorithm respects the field-of-view limits — much like a bird doesn't peer under its own wing. Once the target is centered, a precise closing maneuver engages, steering the drone in a perfect arc, as if plunging into a dive. This approach echoes astronomy’s planet transiting a star, except here the object crosses a control zone amid the flight dynamics.

YOLO, the object detector, processes video so swiftly that the machine’s response outpaces the human eye. Even a smartphone can become the brain of an aerial hunter.

Unlike the relentless machines of ‘Terminator,’ this algorithm is built for rescue: finding the missing, inspecting pipelines, monitoring wildfires. A hunter, but with a noble cause.

Scientists
Ludwig BoltzmannJohn von NeumannJosiah Willard GibbsHenri PoincaréWilliam Borucki
Concepts
drone