Neural network with single hidden layer for air traffic volume prediction in uncontrolled airspace

Abstract

This article presents a model enabling more effcient air traffic management achieved by better data use. Appropriate resource allocation is possible if it is based on a high quality air traffic volume forecast. The proposed approach is inspired by procedures used in flow management in air traffic control. Staff planning in controlled airspace is easier because almost all operations are communicated in the submitted flight plan. Short-term prediction of the number of operations in uncontrolled airspace is a much more challenging task. It is correlated with weather parameters and moreover, it naturally fluctuates throughout the day and the season. The relationship between General Aviation (GA) traffic volume and meteorological conditions were modeled using neural network. The obtained results confirm that it is possible to use the decision support system to plan the number of operational sectors. The described results open a scientific discussion about designing tools predicting air traffic volume in uncontrolled airspace. The accuracy of the model can be improved by processing data from additional sources, but it is associated with a significant increase in
the complexity of the solution.