Predictive Control Methods Of Inverted Pendulum Application Research

Inverted pendulum is an absolutely not stability, high order time, multivariable andstrong coupling nonlinear system as well, and it unifies the robot technology intelligentcontrol and computer control organic, therefore, inverted pendulum can be used as atypical object of control. As new methods of control are emerging, usually used invertedpendulum system to test new control method has a strong ability to control, then to findthe best control method.In the1970s, predictive control theory appeared, after30years of development, it hasbecome one of the most representative advanced control methods. The main applicationobjects of predictive control are constraints and multivariable systems. It also has goodcontrol effect on the unconstrained nonlinear systems and unconstrained lineartime-varying system. The predictive control method to model accuracy is not high, andthis method has good robustness. Therefore, it has been widely used.In this dissertation mainly used generalized predictive control algorithm to control thedouble inverted pendulum system simulation. First, build the mathematical model ofdouble inverted pendulum system by Lagrange method, and linearzed it at equilibriumposition, then we get the balanced state equation, and analyzed the correlated behavior ofthe system. And then, analyzed the characteristic and the sphere of application inpredictive control, and simulation study of generalized predictive control, by comparingwith the simulation curve of LQR, the generalized predictive control can effectivelyimprove rapidity performance of the system. Finally, for the characteristic of generalizedpredictive control, proposed a combination of a fuzzy algorithm with generalizedpredictive control method to control the double inverted pendulum system simulation. Forthe defect of dynamic performance in generalized predictive control, use fuzzy inferenceto generalized predictive compensation of error, make up for defects of informationprocessing in the generalized predictive control. The simulation experiment show thatcompared with the generalized predictive control method solely, add fuzzy algorithmgeneralized predictive control can effectively improve dynamic performance of the system,and has a good robustness and rapidity.