Online Stochastic Prediction of Mid-Flight Aircraft Trajectories

Abstract

Online trajectory prediction is central to the function of air traffic control of improving the flow of air traffic and preventing collisions, particularly considering the ever-increasing number of air travellers. In this paper, we propose an approach to predict the mid-flight trajectory of an aircraft using models learned from historical trajectories. The main idea is based on Hidden Markov Models, representing the location of aircraft as states and weather conditions as observations. Using our approach, one is able to make predictions of future positions of currently mid-flight aircraft for each minute into the future, optionally concatenating these positions to form the remaining predicted trajectory of an aircraft. We evaluated the effectiveness of the proposed approach using a dataset of historical trajectories for flights over the USA. Using prediction accuracy metrics from the aviation domain, we demonstrated that our approach could accurately predict trajectories of mid-flight aircraft, achieving an effectiveness improvement of 26% in horizontal error and 32% in vertical error over baseline models with virtually no loss in prediction efficiency.