A new adaptive neural network control scheme based on feedback linearization technique for nonlinear processes

In this paper, a new discrete-time adaptive control scheme based on feedback linearization technique is proposed to control single-input, single-output (SISO) processes with nonlinear and time-varying characteristics. Given this, an affine model of the process, incorporated in the control scheme, needs to be identified in an on-line manner. For this purpose, an on-line identification approach based on an adaptive neural network with growing and pruning radial basis function (GAP-RBF) structure is used for affine modeling. Also, some desired modifications in the neurons growing and pruning criteria of the original GAP-RBF algorithm has been proposed to enhance its performance in on-line identification. The proposed control scheme is evaluated via a highly nonlinear and time-varying continuous stirred tank reactor (CSTR) benchmark problem. The simulation results show the excellent performance of the developed adaptive control scheme for identification and control of the CSTR process.