Probabilistic Multi-Agent Aircraft Landing Time Prediction

TL;DR

The paper tackles the challenge of uncertain, multi-aircraft landing-time predictions by introducing a probabilistic multi-agent framework that outputs Gaussian distributions for each aircraft's remaining time to land. Built on MAIFormer, the architecture combines air traffic scene embedding, a multi-agent trajectory encoder with masked multivariate and agent attention, and a Gaussian parameter decoder trained via negative log-likelihood. It demonstrates superior accuracy and arrival-sequence consistency on ICN ADS-B terminal-airspace data, while providing calibrated uncertainty estimates and interpretable attention patterns. The work highlights the practical value of uncertainty-aware, interaction-aware predictions for air traffic management and ATCo decision support, and discusses avenues for improvement such as incorporating weather and more flexible distributions.

Abstract

Accurate and reliable aircraft landing time prediction is essential for effective resource allocation in air traffic management. However, the inherent uncertainty of aircraft trajectories and traffic flows poses significant challenges to both prediction accuracy and trustworthiness. Therefore, prediction models should not only provide point estimates of aircraft landing times but also the uncertainties associated with these predictions. Furthermore, aircraft trajectories are frequently influenced by the presence of nearby aircraft through air traffic control interventions such as radar vectoring. Consequently, landing time prediction models must account for multi-agent interactions in the airspace. In this work, we propose a probabilistic multi-agent aircraft landing time prediction framework that provides the landing times of multiple aircraft as distributions. We evaluate the proposed framework using an air traffic surveillance dataset collected from the terminal airspace of the Incheon International Airport in South Korea. The results demonstrate that the proposed model achieves higher prediction accuracy than the baselines and quantifies the associated uncertainties of its outcomes. In addition, the model uncovered underlying patterns in air traffic control through its attention scores, thereby enhancing explainability.

Figures (8)

• Figure 1: Comparison of deterministic (left) and probabilistic (right) predictions in estimating the arrival sequence.
• Figure 2: Overview of the proposed probabilistic multi-agent aircraft landing time prediction framework.
• Figure 3: Illustration of air traffic scene embedding. For illustration, the number of aircraft is set to three, with each aircraft represented by three variates (latitude, longitude, and altitude).
• Figure 4: Illustration of masked multivariate attention and agent attention. The colored arrows denote where attention is applied while the gray arrows indicate where attention is masked.
• Figure 5: Illustration of Gaussian parameter decoder.