10 - Safety and Security

EVALUATION OF METHODS FOR ANOMALY DETECTION IN THE APPROACH PHASE OF OPERATIONAL FLIGHT DATA

R. Hähnel, German Aerospace Centre (DLR), Germany

This paper investigates anomaly detection methods for large-scale, unlabeled flight data from commercial aircraft, focusing on the approach phase. Multiple distance-, forecasting-, and reconstruction-based algorithms are compared using missed approaches as a surrogate benchmark. The study highlights strengths, weaknesses, and complementarities of methods beyond traditional exceedance detection.

View Paper

DOI: 10.71945/icas2026_0291