Bayesian learning of models for estimating uncertainty in alert systems: Application to air traffic conflict avoidance

Abstract

Alert systems detect critical events which can happen in the short term. Uncertainties in data and in the models used for detection cause alert errors. In the case of air traffic control systems such as Short-Term Conflict Alert (STCA), uncertainty increases errors in alerts of separation loss. Statistical methods that are based on analytical assumptions can provide biased estimates of uncertainties. More accurate analysis can be achieved by using Bayesian Model Averaging, which provides estimates of the posterior probability distribution of a prediction. We propose a new approach to estimate the prediction uncertainty, which is based on observations that the uncertainty can be quantified by variance of predicted outcomes. In our approach, predictions for which variances of posterior probabilities are above a given threshold are assigned to be uncertain. To verify our approach we calculate a probability of alert based on the extrapolation of closest point of approach. Using Heathrow airport flight data we found that alerts are often generated under different conditions, variations in which lead to alert detection errors. Achieving 82.1% accuracy of modelling the STCA system, which is a necessary condition for evaluating the uncertainty in prediction, we found that the proposed method is capable of reducing the uncertain component. Comparison with a bootstrap aggregation method has demonstrated a significant reduction of uncertainty in predictions. Realistic estimates of uncertainties will open up new approaches to improving the performance of alert systems.

Review on artificial intelligence techniques for improving representative air traffic management capability

“…Durand et al [93] proposed a neural network based on unsupervised learning, which could calculate almost optimal trajectories, thus solving the problem of two aircraft collision avoidance with the highest reliability when calculating the headings with the resolution of a few milliseconds. Sislak et al [94] presented two agent-based cooperative decentralized aircraft collision avoidance algorithms that worked with different levels of coordination autonomy, making realistic assumptions about the accuracy of flight execution (integrating required navigation performance), where planning interlaced with the planned execution phase. Because the uncertainty in the data and the model used for detection can lead to the TCAS alarm errors, Schetinin et al [95] proposed an uncertainty estimation model for early warning systems based on Bayesian learning.…”

GMDH-Type Neural Networks for Predicting Financial Time Series: A Study of Informational Efficiency of Stock Markets

“…The use of such samplers can additionally provide a prediction with the reliable estimates of predictive posterior density distribution, which is critically important for making risk-aware decisions in the presence of uncertainties, see e.g. [29].…”

“…In particular ML methods have been efficiently used to solve problems such as detection of abnormal patterns and evaluation of brain development in newborn electroencephalograms [30]. The reliable results have been achieved in prediction of trauma survival [28], air-traffic collision avoidance [29], as well as in detection of bone pathologies [13].…”

A domain-specific modeling approach supporting tool-chain development with Bayesian network models

“…Eman et al provided one learning cost-sensitive Bayesian networks for assessing credits for cyber security [26]. Schetinin et al made use of Bayesian network models to support uncertainty estimations for air traffic conflicts [27]. Such developed models provided the realistic insights into predictive distributions for conflicts analysis.…”