Model And Control For Rotary Multistage Inverted Pendulum Based On Fuzzy Inference

Inverted pendulum system is a typical model of multivariable, nonlinearity, strong-coupling and absolutely instability. It is the ideal platform to verify the algorithm. What is more, the research of inverted pendulum has many important realistic meaning in the research. The rotary inverted pendulum system is used as the research object and this paper mainly adopts fuzzy inference model method modeling the rotary double inverted pendulum and rotary triple inverted pendulum. What's more, the control systems of the inverted pendulum are researched. The main works in this paper are the following:Firstly, after concisely introducing the structure and moving characters of the rotary inverted pendulum system, the mechanism and mathematical models of rotary triple inverted pendulum are presented, using Lagrange equation method. Then, the fuzzy systems of the roatry inverted pendulums are modeled by using the marginal linearization method. It is proved that open control system of rotary multistage inverted pendulum system is instability, but it is controllable and observable on equilibrium point.Secondly, linear quadratic optimal controller and robust guaranteed cost controller are designed for rotary inverted pendulums respectively. Under MATLAB/SIMULINK circumstance, controls of fore-mentioned controllers for double and triple inverted pendulum are effective. Some contrast analysis between different controllers is done carefully.Thirdly, the main components and control mechanism of roatry double inverted pendulum hardware systems are briefly introduced. The program of robust control algorithm for double inverted pendulum is designed. When the pendulum is stable, the response curves of numerical control system are obtained with disturbance. In order to test the robustness and the capacity of resisting disturbance, lots of experiments are done and the results are analyzed in very great detail.Simulation result and experiments show that the fuzzy inference systems based on the marginal linearization method are effective and the controllers designed in this paper have a good stability, robustness and immunity.