Koki Ho
Georgia Institute of Technology
This research proposes a new integrated framework for identifying and prioritizing safe landing locations in real time to make inflight divert maneuvers. The state-of-the-art algorithms for landing zone selection only use local terrain features such as slopes and roughness; little research has focused on algorithms for an integrated system-based selection of landing zones. In response to this challenge, we propose a novel deep-learning-based framework to leverage on-ground training to enable the onboard algorithms to quickly detect hazards and identify safe landing zones with inflight divert maneuvers in consideration.