Controladores baseados em técnicas de inteligência computacional: análise, projeto e aplicações

• Ano de defesa: 2007
• Autor(a) principal: Anderson Vasconcelos Pires
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Federal de Minas Gerais; UFMG
• Palavras-chave em Português: Engenharia Elétrica; Engenharia elétrica
• Link de acesso: http://hdl.handle.net/1843/BUOS-8CRE4P

This work introduces a study about On-Line Neurofuzzy Controller (ONFC) through the mathematical analysis of its equations that allow understand it more deeply. Still on the ONFC, is introduced the neural and sliding modes versions of controller. The inverse dynamic control is boarded also where are presented some points that complicate to get theinverse artificial model and what can be done to facilitate the attainment of such model. Because of some limitations of inverse dynamic control and to be a controller that does not need to have a so great knowledge of system to be controlled, is propose the approached inverse dynamic control. In this proposal, the artificial model of the system is used to get the inverse function. The artificial model of the system is easier to get than inverse model. This allows obtaining better results in approached inverse dynamic control, especially in control using on-line training.