EFFICIENCY OF DETERMINATION METHODS OF CONTROLLERS’ PREFERENCES GROUP SYSTEMS BASED ON THE DANGER OF CHARACTERISTIC ERRORS IN THE PROCESS OF AIR TRAFFIC CONTROL

• Main Authors: Алексей Николаевич Рева, Шахин Шахвели-оглы Насиров, Бала Мушгюль-оглы Мирзоев
• Format: Article
• Language: English
• Published: National Aerospace University «Kharkiv Aviation Institute» 2018-12-01
• Series: Авіаційно-космічна техніка та технологія
• Subjects: безопасность полётов; опасность ошибок авиадиспетчеров; системы предпочтений; методы построения; классические критерии принятия решений; медиана кемени; непараметрическая оптимизация
• Online Access: http://nti.khai.edu/ojs/index.php/aktt/article/view/360

The human factor problem should be solved by identifying, qualifying and preventing the erroneous actions of the air traffic controllers. It is presented two schemes explaining the structure of human qualimetry factor and the interaction of the components of the ICAO safety concept, where the main emphasis is on an aviation personnel’ attitude to dangerous actions or conditions, which is revealed by the qualimetry of the decision-making processes’ characteristics: the attitude towards risk (the main dominants and fuzzy assessments), levels of claims, dangerous qualities and preferences systems. The preferences systems are considered as ordered characteristics and indicators of professional activity, which are subjectively compared with the positions of influence on flight safety. The spectrum of n = 21 characteristic errors was formed considering the recommendations of ICAO, EUROCONTROL and accident statistics. It is determined that procedures of collecting the information of errors danger contribute their recognition, memorization, and avoidance: controllers who passed the test according to the proposed method before training made by one third fewer errors in its process. Two criteria for assessing group preferences are realized: the level of consensus (known as Kendall’s coefficient of concordance) and the severity of the ranking, determined by the presence of "related" ranks, for which a special indicator is introduced. It is defined that this indicator should be determined both for the sample of respondents and for the preferences group systems of developed with the chosen method of individual opinions’ aggregation. It was performed the comparative analysis of complex decision-making strategies of effectiveness in the construction of a preferences group systems m = 65 controllers: sum and averaging of ranks, classical criteria (Wald's, Savage's and Laplace's criterion), optimal prediction, applying the non-parametric optimization of the preferences group systems. The non-parametric optimization of the group system of pre-readings was carried out by Kemeny median and it was proved that it was the closest to all the results obtained by other methods and strategies