航空機対比較による東京国際空港到着順予測

抄録

Accurate prediction of aircraft trajectory is indispensable for efficient air traffic management under trajectory-based operation concept. Aircraft are vectored to make an efficient arrival sequence to Tokyo International Airport, and the trajectories of aircraft depend on the order of arrival sequence. Prediction of order of arrival sequence contributes accurate prediction of aircraft trajectories. The author developed a model that predicts the order of arrival sequence to Tokyo International Airport when aircraft pass 50 nautical miles, 100 nautical miles and 150 nautical miles points to the airport. The model includes a support vector machine (SVM) model that predicts which of a pair of aircraft is a leader in the sequence. The accuracy of prediction by our SVM model at 150 nautical miles point is 92.3 percent. The order of arrival sequence is predicted by the integration of the SVM model and quick-sort algorithm. The accuracies of predictions of arrival sequence at 50 nautical miles, 100 nautical miles and 150 nautical miles points are 95 percent, 89 percent and 76 percent, respectively. They are much higher than the accuracies of predictions based on first-come-first-served principle, which are 85 percent, 66 percent and 49 percent, respectively. This shows the effectiveness of our model.