Semiglobally ISpS Disturbance Attenuation via Adaptive Neural Design for a Class of Nonlinear Systems

This paper considers the disturbance attenuation problem in certain practical sense for a class of multi-input multi-output nonlinear systems with structure uncertainties which are modelled by unknown smooth function vectors and matrices. The system under consideration is composed of two cascaded subsystems: a zero-input asymptotically stable nonlinear one and a linearizable one which is controllable. The output of the system is a function vector of the states of these two subsystems. To solve disturbance attenuation problem and cope with the unknown function vectors and matrices in the system, an adaptive control scheme based on neural network technique and Lyapunov theory is developed for the nonlinear system. The resultant closed loop system is semiglobally input-to-state practically stable (ISpS) and the responses of the output to the disturbance inputs are attenuated in certain practical sense. Simulation studies are conducted to verify the effectiveness of the scheme