Event-triggered neural intelligent control for uncertain nonlinear systems with specified-time guaranteed behaviors

In this paper, an event-triggered neural intelligent control for uncertain nonlinear systems with specified-time guaranteed behaviors is proposed. To cope with constrained communication resources, an event-triggered mechanism using switched thresholds is devised without involving input-to-state stability assumption, such that a better design flexibility and freedom can be provided. In addition, a minimum-learning-parameter-based state observer is developed to online estimate the unavailable states and uncertainties at the same time, which effectively eliminates the issue of learning explosion without sacrificing the identification precision. Furthermore, in pursuit of making a compromise between sampling cost and tracking performance, a modified barrier Lyapunov function based on a time-varying finite-time behavior boundary is constructed in the controller design, which can guarantee that the tracking error converges to a predetermined region within a specified time. Then by introducing the Nussbaum gain technique to handle the unknown control direction, an event-triggered neural output feedback control strategy is synthesized within the framework of dynamic surface control. Meanwhile, with the aid of Lyapunov synthesis, all the signals involved in the closed-loop system are proved to be bounded while Zeno phenomena is circumvented, and system outputs are well within the predefined region. Finally, an application on control design for a micro-electro-mechanical system gyroscope is given to validate the efficiency and superiority of proposed intelligent control scheme.

Automated summary

This paper proposes an event-triggered neural intelligent control for uncertain nonlinear systems with guaranteed specified-time behaviors, addressing communication resource constraints and providing design flexibility. A minimum-learning-parameter-based state observer is developed to estimate unavailable states and uncertainties, a modified barrier Lyapunov function is constructed to balance sampling cost and tracking performance, and an event-triggered neural output feedback control strategy is synthesized within the framework of dynamic surface control. The efficiency and superiority of the proposed intelligent control scheme are validated through an application on control design for a micro-electro-mechanical system gyroscope.