A Decision-Making Framework for Autonomous Aerial Vehicles in Responding to GPS Loss

Unmanned Aerial Vehicles (UAVs) face critical challenges when operating in environments where Global Positioning System (GPS) signals are degraded or unavailable. Existing work on GPS-denied navigation has produced capable localization and control techniques, but these contributions are often siloed and rarely integrate energy constraints, decision logic, and regulatory considerations into a cohesive architecture, especially in Beyond-Visual-Line-Of-Sight (BVLOS) operations. This paper proposes a unified, modular framework that couples resilient multimodal localization, such as visual–inertial odometry and simultaneous localization and mapping, with an energy-aware trajectory planner and a contingency-aware decision layer that manages actions such as rerouting, hovering, reorienting, returning to home, or performing a safe landing. The proposed framework explicitly encodes navigation confidence, command-and-control integrity, and logging for post-flight audit, enabling traceable behavior suitable for certification and large-scale deployment. By moving GPS-denied response from ad-hoc failsafe logic to a standardized, closed-loop decision-making model, this work provides a blueprint that can inform future performance standards and accelerate trustworthy beyond-visual-line-of-sight and advanced air mobility operations in real-world industrial, public-safety, and urban contexts.