Analysis and Modeling of Air Traffic Trajectories Uncertainty in Chinese Airspace

The increasing pressure on air traffic management (ATM) system has become a key issue that impedes the development of air transportation. Therefore, a transformation is underway to increase ATM safety, capacity, efficiency and environmental friendliness. As a fundamental element of the transformation, trajectory-based operation (TBO) considers the trajectory during all phases of flight and supports strategic planning to maximize the ATM system capacity. However, it is hard to guarantee the accuracy of trajectory due to the effects of meteorological conditions, airspace adjustments, airport capacity limitations and etc.. Thus, the analysis and modeling of trajectories uncertainty based on real data is proposed to quantify those effects. Firstly, the flight and trajectory data for Chinese airspace within three months are analyzed and the characteristic factors which have great influence on trajectories uncertainty are selected. Then, setting the key characteristic factors as input and the arrival time at the waypoint as output, the supervised learning model is established by SVM and RNN respectively. Finally, the predicted results of the two methods and the real data have been compared, and the accuracy of the core factors and the model have been verified.