A Tight Adaptive Event-Triggered Controller With Global Prescribed Tracking Performance

This article proposes a new adaptive event-triggered controller with prescribed tracking performance for a class of uncertain nonlinear systems. The controller is global, instead of semiglobal, by eliminating the initial-condition-dependence on the performance function. In contrast to the existing results, the controller design is tight to reduce conservatism, in which two enablers are involved. First, the controller fully leverages available information on system nonlinearities, rather than discarding the information as is done in the context of funnel control (FC). This enables the controller to more efficiently counteract the nonlinearities, potentially avoiding unnecessarily large control effort, thereby reducing conservatism. Second, dedicated dynamic compensations are introduced with the help of tuning functions to compensate for the intrinsic uncertainties appearing in the execution error, system nonlinearities, and control coefficients. In this way, estimating conservative bounds of multiple uncertain parameters is circumvented, and especially, inequality estimates, including completing squares, are largely avoided, thereby also reducing conservatism. To further improve communication efficiency, the proposed control scheme is extended to the scenario where the information transmission from sensor to controller is also event-triggered, by delicately designing a double-sided event-triggering mechanism. A classical pendulum system is utilized to verify the effectiveness and superiority of the proposed method.