"Take Me Home, Wi-Fi Drone"
A Drone-based Wireless System for Wilderness Search and Rescue
University of Hong Kong
(This work is accepted by MobiCom 2026! 🎉🎉)
A Drone-based Wireless System for Wilderness Search and Rescue
(This work is accepted by MobiCom 2026! 🎉🎉)
Tree canopy creates visual obstruction, even infrared is difficult to penetrate for identification [2]
Find lost person through Wi-Fi device; identify via auto-connection
Wi²SAR leverages the automatic reconnection behavior of modern Wi-Fi devices. By simulating a known network (e.g., home Wi-Fi), the drone elicits identifiable packets from the victim's device. This mechanism enables rapid discovery and identification of the target victim without requiring any prior app installation or active user cooperation.
A custom 3D-printed Luneburg Lens is integrated to amplify directional signal strength differences. This passive optical device enables a novel RSS-only, long-range 3D Angle-of-Arrival (AoA) estimation method. It significantly extends the operational range and eliminates the need for complex phase calibration on moving drones.
The system employs an adaptive drone navigation scheme that guides the drone toward the target efficiently. It uses a dual-phase search strategy: an initial exploratory grid search to detect the signal, followed by a direction-guided approach that continuously refines the heading based on real-time AoA estimates to pinpoint the victim's location.
Range Extension
Median Angular Error
Discovery Rate
Localization Accuracy