SAMPLED-DATA INDIRECT ADAPTIVE-CONTROL OF BIOREACTOR USING AFFINE RADIAL BASIS FUNCTION NETWORK ARCHITECTURE

This paper considers a problem of bioreactor control, which is formula ted in Anderson and Miller (1990) and Ungar (1990) as a benchmark prob lem for application of neural network-based adaptive control algorithm s. A completely adaptive control of this strongly nonlinear system is achieved with no a priori knowledge of its dynamics. This becomes poss ible thanks to a novel architecture of the controller, which is based on an affine Radial Basis Function network approximation of the sample d-data system mapping. Approximation with such network could be consid ered as a generalization of a standard practice to linearize a nonline ar system about the working regime. As the network is affine in the co ntrol components, it can be inverted with respect to the control vecto r by using fast matrix computations. The considered approach includes several features, recently introduced in some advanced process control algorithms. These features-multirate sampling, on-line adaptation, an d Radial Basis Function approximation of the system nonlinear-are cruc ial for the achieved high performance of the controller.