HUNT: High-Speed UAV Navigation and Tracking in Unstructured Environments Via Instantaneous Relative Frames
Alessandro Saviolo, Jeffrey Mao, Giuseppe Loianno
AI summary
Problem
UAVs struggle to navigate safely at high speeds and track targets in GPS-denied, unstructured environments because traditional global localization degrades without persistent landmarks, while existing relative tracking methods fail when targets are out of view.
Approach
HUNT uses an instantaneous relative frame anchored to directly observable onboard states for safe high-speed loitering, then seamlessly re-anchors to a detected target for tracking, all enforced by a control barrier function-augmented nonlinear model predictive controller.
Key results
- Stable high-speed loitering (4.1 m/s) with bounded heading error in dense forests and urban canyons without GPS
- Seamless, oscillation-free mode switching between loitering and tracking via confidence-based covariance thresholds
- Robust acquisition and pursuit of dynamic targets in cluttered, GPS-denied environments where global methods fail
- Real-time collision avoidance guaranteed by embedding control barrier functions directly into the NMPC
Why it matters
Enables reliable, high-speed autonomous search-and-rescue operations in GPS-denied, unstructured environments where traditional localization fails.
Abstract
Search and rescue operations require unmanned aerial vehicles to both traverse unknown unstructured environ- ments at high speed and track targets once detected. Achieving both capabilities under degraded sensing and without global localization remains an open challenge. Recent works on relative navigation have shown robust tracking by anchoring planning and control to a visible detected object, but cannot address navigation when no target is in the field of view. We present HUNT (High-speed UAV Navigation and Tracking), a real- time framework that unifies traversal, acquisition, and tracking within a single relative formulation. HUNT defines navigation objectives directly from onboard instantaneous observables such as attitude, altitude, and velocity, enabling reactive high- speed flight during search. Once a target is detected, the same perception–control pipeline transitions seamlessly to tracking. Outdoor experiments in dense forests, container compounds, and search-and-rescue operations with vehicles and mannequins demonstrate robust autonomy where global methods fail. Video: https://youtu.be/YsSflqPPHhs