Kolesnyk A. Methods of improving the efficiency of air traffic services in the event of emergency situations in flight

The dissertation develops a model for assessing the risk of emergency flight situations on the example of engine failure based on the Bayesian network. This model allows to calculate the probabilities of consequences of the specified danger, to determine the most probable consequence of the danger and the level of risk based on the ICAO classification and will increase the validity of the decision of the air traffic controller in emergency flight situations. As a decision-making criterion in the dissertation the criterion of minimization of risk of occurrence of undesirable consequences as a result of realization by the air traffic controller of the accepted decision is offered A method for determining the alternative to the end of the flight in case of emergency situations in flight has been developed. Within the framework of the developed method, based on Bayesian networks, the following models have been built: a model for determining the possibility of continuing the flight, models for determining potential damage during landing at the aerodrome and the site. Due to the consideration in the built models of certain groups of factors that affect the outcome of the decision at each stage, the developed method will increase the validity of the decision of the aviation dispatcher in off-duty flight situations. A method for automating the decision-making process by an air traffic controller for special cases in flight has been developed. The method is based on probabilistic programming technology. The developed method involves building a probabilistic hierarchical model of an emergency situation based on the Bayesian network, assigning conditional and unconditional probabilities to the vertices of the network, creating a sample of learning examples to teach the created model to improve the quality of predicting future situations, choice of a model learning paradigm. The application of the developed method will increase the efficiency and validity of the decision of the aviation dispatcher in emergency flight situations.