Research On The Application Of Fuzzy Neural Network Control In Double Inverted Pendulum System

Inverted pendulum is a typical model of multi-variable,nonlinear, essentially unsteady system,and researching stability of inverted pendulum system has the profound meaning in theory and methodology.The research on inverted pendulum can be diverted to the research on nonlinear,multi-variable and unsteady system. In this dissertation, first of all,analyze the meaning of researching the inverted pendulum system,give a summary on the research actuality of inverted pendulum,and introduce many control ways on making inverted pendulum system steady.Based on the summary of fuzzy control theory and neural network control theory,the merit of combination fuzzy logic control with neural network control is presented. Then introduce the essential definitions, analyze the structure of fuzzy neural network, put forward a new type fuzzy neural network,and analyze the characteristic of approximation to arbitrary nonlinear function, which makes the fuzzy neural network model approximate any physical nonlinear system within arbitrary accuracy.Analyze fuzzy clustering algorithm and fuzzy c-means, give a method of dynamic fuzzy model identification for multi-variable system,research the stability of a continuous system and discrete-time system based on dynamic fuzzy model,and present the methods of designs for the multi-Variable fuzzy neural network controller for the continuous systems and the discrete-time systems.Introduce the structure of the double inverted pendulum system,establish its model and deduces the state-space expression.Analyze the stability,controllability and observability and design the linear regulator. Fuzzy neural network is combined with fuzzy clustering to determine dynamic fuzzy model of the double inverted pendulum,and give a design method for fuzzy neural network system controller.Finally,a simulation of double inverted pendulum system demonstrates the performance of the model and controller.