An Integrated Clustering-based Approach for Aircraft Trajectory Generation Model

Accurate and efficient four-dimensional trajectory (4DT) calculation is crucial for Trajectory Based Operation (TBO) of air transportation. In order to create precise 4DT and provide a reference for the fine-grained Air Traffic Management(ATM), an integrated trajectory clustering for trajectory generation framework is proposed mining the main trajectory patterns and realizing the large-scale 4DT generation fast. Firstly, DBSCAN and TraClus clustering methods are accommodated into one integrated procedure to extract both the global and local features of trajectory. Then, some typical trajectory points mined above are adopted as the origin and destination waypoints to generate the target air-route trajectory via a directed network with transition probability. Two trajectory generation strategies are presented further to obtain the most possible trajectory and its related feasible trajectory set. Finally, through experiments with trajectory data collected on air-route from Beijing to Guangzhou, China, both the validity and accuracy of the proposed method are verified.