Neural network model based indirect sliding mode controller design

This paper describes a unified framework for designing a nonlinear controller for a plant which is known to be nonlinear, yet for which no appropriate model is available for nonlinear controller design. The indirect sliding mode approach is exploited for controller design. This method uses sliding mode techniques to effect asymptotic linearisation of a nonlinear system expressed in generalised controller canonical form. It is shown that neural networks can be exploited to generate such a nonlinear model. The effectiveness of the proposed scheme is illustrated using a design example.