Classificação de eventos de PIO utilizando Redes Neurais Artificiais

Abstract

During the certification and the flight tests of a new airplane, the verification process to prove the low susceptibility at PIO is usually done for the experience of the test pilots, thus, the result and safety operation regarding PIO is intrinsic the assessment subjectivity, in other words, the certification process of PIO is directly related with the test pilots experiences. New airplanes tend to reduce stability margins, that is, increasing PIO tendencies, and full autonomous airplanes with capacity to perform a complete flight without crew, for this reason it is necessary the reduction of the subjectivity of PIO assessment for next generation of aircraft. This project covers the development of a neural network that will be able to classify the PIO oscillations in accordance with the PIOR Scale and to the subjective assessment of PIO classification.Neural networks assertiveness was higher for PIO classification than the result done by pilots, when the choice of inputs follow the guidelines established for this work.