On the Use of Generative Adversarial Networks for Aircraft Trajectory Generation and Atypical Approach Detection

Aircraft approach flight path safety management provides procedures that guide the aircraft to intercept final axis and runway slope before landing. In order detect atypical behavior, this paper explores use of data generative models learn real probability distributions identify flights do not follow these distributions. Through Generative Adversarial Networks (GAN), a GAN is first trained paths, generating new from learned Experiments show generated realistic patterns. Unlike trajectories by physical models, proposed technique, only based on past data, able account for external factors such as Air Traffic Control (ATC) orders, pilot behavior or meteorological phenomena. Next, used abnormal compare results with clustering technique combined functional principal component analysis. The reported non compliant are relevant.