An improved trajectory prediction algorithm based on trajectory data mining for air traffic management

Trajectory prediction is an important technology for ensuring safety and efficiency of the air traffic. Hybrid estimation algorithm and intent inference algorithm are usually used to make long-term probabilistic trajectory prediction. In this paper, data mining algorithms are used to process the historical radar data and to abstract a typical trajectory library. An improved trajectory prediction algorithm is proposed based on the typical trajectory, which is used as the intent information to update the transition probability matrix, and is also used to propagate the nominal trajectory instead of the flight plan path. The prediction performance of the proposed algorithm is tested using real radar data from North China Air Traffic Management Bureau. The simulation results show that the improved algorithm has a better prediction performance and the prediction accuracy is improved by 10% at most.