Abstract
The double inverted pendulum has been known as a good example of nonlinear system. In the double inverted pendulum there are four equilibrium points. In this paper, we present a control method for transition between equilibrium points of the double inverted pendulum. The swing-up controller is designed based on energy. The stabilizing controllers are designed on the basis of the linear models at each unstable equilibrium point. Radial Basis Function Neural Network (RBFNN) learned by genetic algorithm is used to integrate the swing-up controller and the stabilizing controllers. The effectiveness of the proposed method is shown by simulation and experimental results.
