Real-time Implementation of a Neuro-Fuzzy Control System for the Inverted Pendulum

• Autores: Rojas-Moreno, Arturo, Nuñez-Ocola, César, Merchan-Gordillo, Fernando, Córdova-Sosa, Luis E.
• Formato: artículo
• Fecha de Publicación: 2001
• Institución: Universidad Nacional de Ingeniería
• Repositorio: Revistas - Universidad Nacional de Ingeniería
• Lenguaje: español
• OAI Identifier: oai:oai:revistas.uni.edu.pe:article/518
• Enlace del recurso: https://revistas.uni.edu.pe/index.php/tecnia/article/view/518
• Nivel de acceso: acceso abierto
• Materia: Inteligencia Artificial; Diseño; Controladores; Lógica Difusa; Takagi-Kosko-Sugeno; Neuro-Difuso; ANFIS, Péndulo Invertido; Equilibrio de una Varilla; Carro; Péndulo; Artificial Intelligence; Design; Controllers; Fuzzy Logic; Neuro-Fuzzy; ANFIS, Inverted Pendulum; Balance of a Rod; Cart; Pendulum

This article presents the design and implementation procedure of a TKS (Takagi-Kosko-Sugeno) fuzzy controller that uses the ANFIS (Adaptive Neuro-Fuzzy Inference System) tuning technique. The developed methodology is applied to a classic case study, but with particularly attractive characteristics to test any type of controller: the Inverted Pendulum System (INS). The SPI comprises a rod attached at one end to a pivot mounted on a carriage that moves in a straight line on rails. The objectiveControl is to keep the rod in a vertical position by means of a force applied to the carriage. The design procedure requires: modeling the SPI, selecting a control strategy according to the input/output variables of the model, selecting the control strategy according to the types of control loops on the variables, and, the design of the fuzzy controller using ANFIS to optimize its performance. After the system has been simulated, the real-time implementation is carried out using a HW/SW (Hardware/Software) interface. The control software is written in C++. The experimental results obtained validate the developed design procedure.