| Popis: |
A fuzzy neural network model of the air traffic control information system has been developed. When constructing a fuzzy neural network model of the air traffic control process, it is necessary to study the reliability of each flight crew member's work on the formation of aircraft flows in their area of responsibility, including the occurrence of abnormal situations in the air and on the ground. In this case, the construction of the model requires means of processing, formalization of fuzzy information. Thus, the construction of this model requires the use of a formal apparatus of fuzzy algebra, fuzzy logic, fuzzy identification theory, that is, methods for constructing fuzzy models based on the results of observations. To describe the elements of the fuzzy neural system in the process of building the model, the concepts of fuzzy set, linguistic variable are used and the construction of the membership function of fuzzy sets is performed. Given that within the fuzzy zone of each term the law of change of each variable is clearly nonlinear, the form of membership functions was chosen by Gaussian and sigmoid form. Expert assessments of specialists are used to build training samples. When drawing up the rules of fuzzy products are also taken into account in expert assessments, the relationship and significance of factors relative to each other in each block. The output of each module determines the efficiency of a particular process. It depends not only on the speed of the process itself, but also on the quality of air traffic control in a particular module (for example, whether the intervals of horizontal and vertical separation have been violated), which ultimately affect the safety and regularity of air traffic. Thus, the formalized experience of experts allowed to build a fuzzy neural network model of air traffic control information system. The model includes six modules that implement, respectively, one of the schemes for the formation of aircraft flows using a fuzzy neural network. The seventh module of the model is intended for previous modules. The information model of air traffic control has been further developed, which, unlike the known ones, is a hierarchical neural network with a vague description of the factors influencing the air traffic control process. The information model allows to increase flight safety. |
