Aplicação de Sistemas Neuro-Fuzzy no Controle de Aeronaves em Operações Críticas de Voo

• Ano de defesa: 2021
• Autor(a) principal: Pereira, Bruno Luiz
• Tipo de documento: Tese
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Federal de Uberlândia; Brasil; Programa de Pós-graduação em Engenharia Mecânica
• Palavras-chave em Português: Análise Individual Ponderada; Pondered Individual Analysis; Controle de Aeronaves; Aircraft Control; Controle Inteligente; Intelligent Control; Controle Neuro-Fuzzy; Neuro-Fuzzy Control; LOC-I; Método de Inferência Fuzzy; Fuzzy Inference Method; PIA; SBRF; FRBS; Sistema Neuro-Fuzzy; Neuro-Fuzzy System; Teoria dos Conjuntos Fuzzy; Fuzzy Sets Theory; CNPQ::ENGENHARIAS::ENGENHARIA MECANICA; CNPQ::ENGENHARIAS::ENGENHARIA MECANICA::MECANICA DOS SOLIDOS; CNPQ::ENGENHARIAS::ENGENHARIA ELETRICA::ELETRONICA INDUSTRIAL, SISTEMAS E CONTROLES ELETRONICOS::CONTROLE DE PROCESSOS ELETRONICOS, RETROALIMENTACAO; CNPQ::ENGENHARIAS::ENGENHARIA AEROESPACIAL::DINAMICA DE VOO::ESTABILIDADE E CONTROLE; CNPQ::CIENCIAS EXATAS E DA TERRA::MATEMATICA::MATEMATICA APLICADA::ANALISE NUMERICA; CNPQ::CIENCIAS EXATAS E DA TERRA::CIENCIA DA COMPUTACAO::MATEMATICA DA COMPUTACAO::MODELOS ANALITICOS E DE SIMULACAO; Engenharia mecânica; Voo - Controle; Aeronautica - Engenharia; Aeronaves - Segurança
• Link de acesso: https://repositorio.ufu.br/handle/123456789/33987; http://doi.org/10.14393/ufu.te.2021.555

Loss of control in-flight is the cause of approximately 70% of all fatalities occurring in aircraft with a take-off mass greater than 5,700 kg. The relevance of this subject provokes the technical and scientific community, and leads to a series of discussions and the generation of norms, procedures and devices that seek to mitigate the causes of these accidents. In order to contribute to the aeronautical sector with regard to the development of new strategies that seek to minimize the number of fatal air accidents, this work proposes the use of a new control architecture based on the combination of neuro-fuzzy systems in the control of aircraft in critical flight operations. For that matter, a new fuzzy inference method, called PIA (Pondered Individual Analysis), is developed, which combines intuitiveness and high computational performance in the process of mathematical translation of the rule base involved in the process. The validation of the proposed technique involves the development of a software-inthe- loop simulation between MATLAB and X-Plane 11, in which the ability of the proposed control architecture, in critical flight operations, to maintain the response of the aircraft around the reference signals is verified, and also it is analyzed the performance of the control system when taking into account a dynamic model raised from experimental data, extracted from a flight test carried out on a small scale Cessna 172 aircraft. The results of the longitudinal and lateral-directional dynamics of the aircraft are analyzed and compared to those obtained with the proportional integral derivative controller and with the neuro-fuzzy controller that uses theTakagi-Sugeno fuzzy inference method, and they present lower mean absolute error in relation to the desired behavior for the aircraft, and thus highlight that the PIA method demonstrates to be an effective tool to be considered in solving problems in the control area.