Sintonia do controle de configuração de robôs móveis multiarticulados via algoritmo genético

• Ano de defesa: 2013
• Autor(a) principal: Bertolani, Diego Nunes
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Federal do Espírito Santo; BR; Mestrado em Engenharia Elétrica; Centro Tecnológico; UFES; Programa de Pós-Graduação em Engenharia Elétrica
• Palavras-chave em Português: Genetic algorithms; Multiarticulated mobile robots; Tuning; Robôs móveis multiarticulados; Sintonia; Algorítmos genéticos; Robôs - Sistemas de controle; Robôs móveis; Sistemas inteligentes de controle; Eletrônica Industrial, Sistemas e Controles Eletrônicos; 621.3
• Link de acesso: http://repositorio.ufes.br/handle/10/6205

This work proposes a strategy for tuning linear controllers via genetic algorithms in the space of configurations of Multiarticulated Mobile Robots to the problem of backward motion control. As in any nonlinear system and complex, the linear closed-loop control is necessary to provide robustness to the system nonlinear controllers or approximate left inverse of the plant in open loop. For this problem, the nonlinear controllers proposed in the literature have gains of linear controllers empirically adjusted how constant values for movements that cover a wide range of values of the angle configuration, always aiming to avoid the situation of pileup or jacknife composition. This approach has led to results in transitional or stationary unsatisfactory. Thus, this work is done a systematic study of tuning gains in around appropriate partitions of the configuration space, to identify the variability of earnings due to the best performance obtained in each partition. The study is based on systematic experimental two nonlinear controllers proposed for a robot or multiarticulated vehicle scale, consisting of one element tractor and two trailers liabilities or without motorization. The proposed strategy has how the basic tool for implementing an developed interface to allow versatility in the analysis of various linear structures with multiple adjustable gains, various control structures, and various nonlinear plants made possible through the analytical model or numerical approximations, for example, neural, and fuzzy. Beyond the scope of this study, the identification of vector gains for several reversals of partitions configuration space shouldresult in the synthesis of a further cascade element in the overall control in the form of an interpolation of gains, possibly fuzzy which should provide a satisfactory performance in the execution of more complex maneuvers that require larger movements, fast and accurate.