| Inverted pendulum system is a typical multi-variable, non-linear, strong-coupling, rapid-moved and natural-unstable system. With the development of intelligent control, the new control methods will continue to emerge, and people can test the control effectiveness of these new methods by inverted pendulum system, which is a strict control object.This paper establishes the mathematics models of the inverted pendulum, and does some analysis about the stability and controllability of them; researches the adaptive neural network fuzzy inference system theory, and brings the"refining discrete state space point network" strategy up, which can improve control accuracy wholly; brings the "non-uniform division of the variable interval" strategy up to select separate points, considering the relationship between the nonlinear characteristics of inverted pendulum and its poles'angles, which can improve control accuracy when its poles'angles are deviating from the equilibrium position more greatly; by using above strategies and linear quadratic regulator design theory, designs the sugeno adaptive neural network fuzzy controllers, and researches the control performance in Simulink platform and real inverted pendulum system. The result shows that this control algorithm has better control quality comparing with the linear quadratic regulator control. |
PDF Full Text Request