Radial basis function networks for internal model control

Abstract

In this paper radial basis function (RBF) networks are used in the framework of nonlinear internal model control (IMC). A nonlinear IMC strategy is proposed that includes explicit input weighting. This strategy yields a control law in form of an analytical expression if a control-linear RBF model is used. Important implementation issues such as disturbance modeling and state estimation are discussed and an extended Kalman filter approach is proposed for combined state and model parameter estimation. Simulation studies for a continuous stirred tank reactor with multiple steady states indicate that the proposed control strategy is well suited for the control of unstable nonlinear systems.