Output feedback control of a class of nonlinear systems using direct adaptive fuzzy controller

A stable adaptive fuzzy-logic controller based on output measurements is proposed for feedback linearisable nonlinear systems with unknown dynamics. A direct adaptive controller is derived. The fuzzy system approximates the optimal controller. A state feedback version of the controller is presented first which incorporates a supervisory control term to ensure system stability. The supervisory control technique is shown to be inadequate to provide system stability in output feedback. An output feedback version of the controller is presented which incorporates a high-gain observer with the advantage of casting the problem in a standard singular perturbation form, which helps in stability study. The system is shown to be semiglobally stable without the need for the supervisory control term. A detailed stability analysis is given which shows the boundedness of all system signals. Also, an upper bound on the tracking error norm is derived. Simulation results for a benchmark problem confirm the analysis and illustrate the performance of the proposed controller.