Adaptive tracking control of a class of nonlinear systems using CMAC network

In this paper, an adaptive control based on a Cerebellar Model Articulation Controller (CMAC) network is derived to solve the output tracking problem for a class of nonlinear systems with unknown structured nonlinearities. Without requiring a priori knowledge of the system parameter values, the proposed adaptive control consists of the conventional sliding control and a feedforward compensation with the CMAC network. The sliding control is used as a classical tracking controller for the nominal system and the CMAC network is used to compensate the parametrization errors. It is shown by the Lyapunov approach that the outputs of the closed-loop system asymptotically track the desired output trajectories. The effectiveness of the proposed control scheme is verified with an application to a two degree-of-freedom (DOF) robotic manipulator.