Uncertainty-Aware Automatic Flight Terrain Avoidance System

Aircraft-terrain collision avoidance is a cornerstone of aviation safety, critical to mitigating catastrophic risks posed by unintended ground contact during flight operations. Generally, existing systems rely on simplified models, deterministic assumptions, or a single predefined maneuver. These approaches often neglect dynamic uncertainties, pilot decision-making variability, and real-time computational constraints, limiting their adaptability in complex operational environments. This paper introduces an uncertainty-aware framework for terrain collision prevention that integrates high-fidelity 6-degree-of-freedom flight dynamics with probabilistic maneuver planning. The proposed system employs polynomial chaos expansion to propagate uncertainties, enabling real-time trajectory prediction and collision risk assessment. Combining uncertainty quantification, high-fidelity dynamics, and probabilistic collision detection, this work advances the robustness of autonomous terrain avoidance systems, particularly in high-stress or dynamically complex environments.