Control of a second order system based on neural networks

• Autores: Borja, Mario, Molero, Rudolph, Cuellar, Nilton, Montes, Martin, Separovich, Drago
• Formato: artículo
• Fecha de Publicación: 2008
• Institución: Universidad Nacional de Ingeniería
• Repositorio: Revistas - Universidad Nacional de Ingeniería
• Lenguaje: español
• OAI Identifier: oai:oai:revistas.uni.edu.pe:article/355
• Enlace del recurso: https://revistas.uni.edu.pe/index.php/tecnia/article/view/355
• Nivel de acceso: acceso abierto
• Materia: neurocontrolador; control con redes neuronales; neurocontrolador por refuerzo; controlador neuronal; neurocontroller; control with neural networks; booster neurocontroller; neural controller

The present work shows the simulation and implementation of a "Neurocontroller" in a second order plant. The neural controller, also known as Neurocontroller, was implemented with a multilayer network, where the backpropagation of the error was developed through the "Backprogation" algorithm. The multilayer network, composed of a hidden layer and an output layer, was first simulated in Matlab to get the variation parameters, then it was simulated in Visual C++ to achieve the optimization. The architecture of this multilayer network was varied many times until reaching an optimal form that will be shown as the final architecture. Next, the simulation was done in LabVIEW 8.4, corroborating the simulations in Visual C++. Finally, the neural controller developed in LabVIEW was tested in real time, showing gratifying results and verifying its effectiveness despite simultaneous changes in the parameters.