Safe2Ditch: Emergency Landing for Small Unmanned Aircraft Systems

Abstract

Emergency landing is a critical safety feature for the expected increase of autonomous vehicles in the airspace. This paper presents Safe2Ditch, an autonomous crash management system for landing small unmanned aerial vehicles in populated environments. Using a prepopulated database, the highest-rated landing site is selected. As the vehicle progresses toward the selected landing site, a camera is used to image the area around the landing site. Using a recently developed visual multiple target tracker, moving ground obstacles are tracked and geolocated using a flat Earth model. These inertial target estimates are used to revise the landing site and allow the vehicle to land quickly in the presence of non-cooperative obstacles. The complete system is demonstrated through 2000 Monte Carlo simulations and 16 hardware flight tests.