RAVEN: Resilient Aerial Navigation via Open-Set Semantic Memory and Behavior Adaptation

RAVEN is a 3D open-set memory-based behavior tree framework for aerial semantic outdoor navigation. RAVEN not only navigates reliably toward detected targets, but also performs long-range reasoning to plan toward distant cues and continues informed search even when direct visual evidence is limited. It further supports multi-class object search and on-the-fly task switching within a mission.