Adaptive neural-network predictive control for nonminimum-phase systems

An adaptive neural-network predictive control strategy for a class of nonlinear processes, which exhibit input multiplicities and change in the sign of steady-state gains, is presented. According to the graphic-based determination for neural network architecture associated with prescribed input/output patterns, the feedforward neural network (FNN) is used to capture dynamic and steady-state characteristics of minimum-phase modes over a specified operating range. A one-step-ahead neural prediction algorithm with respect to physical constraints can carry out the offset free performance. Closed-loop simulations demonstrate the effectiveness of the proposed approaches.