Research On Intelligent Control Based On Rotary Inverted Pendulum

As a complicated, high speed, non-linear, multivariable, strongly coupled, naturally unstable non-minimum phase system, inverted pendulum represents the under-driven control family. In the study of many control methods, researchers always lack proper controlled objects, thus lose the opportunity to further the research work and lead to be an obstruction in converting the theoretic results into practical results. Inverted pendulum is a kind of ideal controlled object platform with its simple structure and relatively low cost. In the family of inverted pendulum, rotary inverted pendulum stands out as an effective testing equipment in the experiment and research of control strategies. Under the help of it, various abstract control concepts such as stability, controllability, convergence speed and noise resistance can be displayed conspicuously.This paper analytically investigates the structure of rotary inverted pendulum and the latest relative research productions at home and abroad. On this basis, the non-linear mathematical model of the rotary inverted pendulum system is formulated by Lagrange method and the corresponding control methods are studied. After analyzing the non-linear characteristics of the system, in the vicinity of the equilibrium point of it, the model is linearized, and the state equation is deduced. Combine the LQR control and fuzzy control, make them coordinate with each other, and give full play of the positive points in order to further improve the stability of the pendulum. The result shows that the double structure control scheme makes the pendulum more stable and more precision. A BP neural network is designed, after learning the control samples, it can approach the first-level rotary inverted pendulum controller. The result of the simulation indicates that after select proper number of nerve sells, parameters of the model and samples one-hidden layer BP-NN can approach first-level rotary inverted pendulum controller successfully. Discuss the method of combine fuzzy control and neural network control, bring their strengths into full play, achieve the rotary inverted pendulum control.