Neural networks based adaptive tracking for nonlinear systems

Adaptive Inverse Control (AIC) is a very significant approach for control of unknown linear and nonlinear plants; neural networks based Adaptive Inverse Control (AIC), for unknown dynamical systems has received much attention in recent years due to its generalized and acquiescent characteristics. In this paper a new Radial Basis Function Neural Networks (RBFNN) based technique for desired tracking performance of a nonlinear plant is presented. In this scheme the tracking error is passed through the estimated jacobian of the plant: and then used for updating the parameters of inverse controller. The presented scheme is authenticated through a simulation on a nonlinear plant model of a heat exchanger. The results demonstrate good quality tracking execution and error convergence is attained both in the presence of disturbance and without disturbance in the plant. Also the presented scheme is capable of minimizing the effect of disturbance in the plant.