Uma contribuição metodológica ao projeto, modelagem matemática e ao controle baseado em inteligência artificial de uma cadeira de rodas servoacionada

Abstract

There are many devices designed by humans, which have high nonlinear characteristics that are inherently unstable and variable in time. These devices are difficult to be modeled and controlled. Classics methods that are used for linear systems have limitations when applied to systems with these characteristics. There are modern alternative approaches to studying, modeling and control such systems, but some are called unconventional, better known as knowledge-based and include the use of artificial intelligence tools. This paper describes the development of a model and control system with the aforementioned characteristics, based on a non - conventional control, with the use of fuzzy logic. Fuzzy logic and its inherently low cust (designing and processing) is situated on a level that allows a layman (physiotherapists, orthopedic technicians, doctors, bio - engineers, etc.) to design and execute a project, since it is aware of its usual way. This work investigates a new technic applied in a real case of a quadriplegics wheelchair with tilt option. We present here additionally two simple studies, but that were fundamental in control by the quadriplegic as voice control based on artificial neural networks and another that brings some innovation to the case of the Center of Gravity (CG) of a wheelchair, computed in real time and which differentiates it from the usual techniques in stabilizing devices like this.