ADAPTIVE CONTROL OF A CLASS OF UNKNOWN MULTIVARIABLE NONLINEAR SYSTEM BASED ON DYNAMICAL NEURAL NETWORKS

In this paper a direct adaptive tracking control scheme for a class of unknown multivariable nonlinear system with modeling errors using dynamical neural networks is presented. A stable weight learning algorithm is determined using Lyapunov theory, avoiding iterative training procedures. The feature of this approach is that neither off-line learning phase-nor initial parameter errors small enough are needed. The robust stability of the closed-loop system is proved, with the tracking error being proportional to the magnitude of the modeling error. Simulation results are given to verify the effectiveness of the newly proposed dynamical neural networks-based adaptive control algorithm.