Cleaned ADS-B Trajectories (LHR-ZRH 2025) for AI Training

This dataset provides a fully sanitized, machine-learning-ready corpus of 2,465 commercial flight trajectories for the London (LHR) to Zurich (ZRH) corridor in 2025. Extracted from the OpenSky Network, the raw ADS-B telemetry was subjected to a rigorous 5-step "Physics-Gate" cleaning pipeline (Haversine corridor filtering, 5-second rolling median smoothing, anomaly excision, BADA 3.16 aerodynamic envelope validation, and a -48m geoid correction). It is designed to train Deep Learning models (LSTM) for 4D trajectory optimization.