Globally Intelligent Adaptive Finite-/Fixed- Time Tracking Control for Strict-Feedback Nonlinear Systems via Composite Learning Approaches

This article focuses on the globally composite adaptive law-based intelligent finite-/fixed- time (FnT/FxT) tracking control issue for a family of uncertain strict-feedback nonlinear systems. First, intelligent approximators with new composite updating laws are developed to model uncertain nonlinear terms, which encompass prediction errors to enhance intelligent approximators' learning behaviors and fewer online learning parameters to diminish computational burden. Then, a novel smooth switching function coupled with robust controllers is designed to pull system states back when the transients are out of the approximators' active domain. After that, a modified FnT/FxT backstepping technique is constructed to render output to follow the reference trajectory, and an adaptive law is employed to alleviate the impact of external disturbances. Moreover, input nonlinearities are considered in the design of control laws. In view of the known actuator magnitude and rate saturations, a novel adaptive fast FxT auxiliary variable system is established to effectively diminish saturation time, and further alleviate the impact of the auxiliary signal on the tracking error. Furthermore, when the input is subject to unknown magnitude saturation, rate saturation and dead-zone simultaneously, a smooth function is introduced to approximate the non-smooth input nonlinearities, where the unknown parameters are modeled via the presented global intelligent approximators. It is theoretically confirmed that the proposed control strategies ensure globally FnT/FxT boundedness of all the closed-loop variables. Finally, the validity of theoretical results is testified via a simulation case.