Machine learning approach to predict aircraft boarding

Abstract

Reliable and predictable ground operations are essential for punctual air traffic movements. Uncertainties in the airborne phase have significantly less impact on flight punctuality than deviations in aircraft ground operations. The ground trajectory of an aircraft primarily consists of the handling processes at the stand, defined as the aircraft turnaround, which are mainly controlled by operational experts. Only the aircraft boarding, which is on the critical path of the turnaround, is driven by the passengers' experience and willingness or ability to follow the proposed procedures. We used a recurrent neural network approach to predict the progress of a running boarding event. In particular, we implemented and trained the Long Short-Term Memory model. Since no operational data of the specific passenger behavior is available, we used a reliable, validated boarding simulation environment to provide data about the aircraft boarding events. First predictions show that uni-variate input (seat load progress) produces insufficient results, so we consider expected passenger interactions in the aircraft cabin as well. These interactions are aggregated to a prior-developed complexity metric and allow an efficient evaluation of the current boarding progress. With this multi-variate input, our Long Short-Term Memory model achieves appropriate prediction results for the boarding progress.