Data-Driven Learning Fuzzy Output-Feedback Control with Prescribed Performance for Nonlinear Systems

In this work, a data-driven learning (DDL) fuzzy output-feedback control scheme is constructed for a class of strict-feedback nonlinear systems with fully unknown nonlinearities and unmeasurable states. By utilizing fuzzy logic systems to approximate unknown nonlinear functions, an observer-based DDL output-feedback fuzzy control law is proposed using instantaneous data and historical data concurrently for adaptation. Besides, to address the asymmetric time-varying error constraints, a performance-adjustable prescribed performance function is employed such that the error convergence rate can be specified arbitrarily. With the proposed control strategy, it is shown that all the signals in the closed-loop system are semi-globally uniformly ultimately bounded, with the tracking error converging to a small residual set while allowing for a freely adjustable prescribed performance bound. Finally, simulations on a robotic unmanned surface vehicle validate the efficacy of the proposed method.