Control for nonlinear chaos based on radial basis function neural networks

A method for control of nonlinear chaotic dynamical systems based on radial basis function neural networks is presented. Combining input-output data obtained from a perturbation parameter model with a linear learning algorithm, neural networks are trained to generate the small disturbance control, then to stabilize the chaotic system. The unstable periodic orbit in the Henon map is directed to a stable fixed point by the method. The simulations show the proposed scheme has great effectiveness.