Predictive control of a mixing tank using radial basis function networks

This paper presents a predictive control strategy employing a radial basis function network (RBFN) for the process model. Similarly to other neural network architectures, the RBFN can model any continuous bounded nonlinear mapping, but has the advantage of faster training. This facilitates its use in adaptive situations. The effectiveness of the controller described here is demonstrated using a mixing tank application. In addition proofs are stated regarding the robustness and stability of the closed-loop system