Formal methodology for analyzing the dynamic behavior of nonlinear systems using fuzzy logic

Abstract

Having the ability to analyze a system from a dynamic point of view can be very useful in many circumstances (industrial systems, biological, economical, . . .). The dynamic analysis of a system allows to understand its behavior and response to different inputs, open loop stability, both locally and globally, or if it is affected by nonlinear phenomena, such as limit cycles, or bifurcations, among others. If the system is unknown or its dynamic is complex enough to obtain its mathematical model, in principle it would not be possible to make a formal dynamic analysis of the system. In these cases, fuzzy logic, and more specifically fuzzy TS models is presented as a powerful tool for analysis and design. The TS fuzzy models are universal approximators both of a function and its derivative, so it allows modeling highly nonlinear systems based on input/output data. Since a fuzzy model is a mathematical model formally speaking, it is possible to study the dynamic aspects of the real system that it models such as in the theory of nonlinear control. This article describes a methodology for obtaining the equilibrium states of a generic nonlinear system, the exact linearization of a completely general fuzzy model, and the use of the Poincaré’s methodology for the study of periodic orbits in fuzzy models. From this information it is possible to study the local stability of the equilibrium states, the dynamics of the system in its environment, and the presence of oscillations, yielding valuable information on the dynamic behavior of the system.

Keywords

• Análisis dinámico
• estabilidad
• estado de equilibrio
• linealización
• metodología de Poincaré
• modelado borroso
• sistemas dinámicos
• Takagi-Sugeno (TS) model