Flexible Performance-Based Neural Network Control for Mechanical Systems Under DoS Attack and Event Triggering

In this work, the flexible performance-based control scheme is presented for the networked uncertain mechanical system under the quasi-periodic denial-of-service (DoS) attack and the appointed event-triggering (ET) mechanism. This scheme features the capability of releasing and recovering the prescribed performance in light of the activation and sleeping of the DoS attack. These functionalities attribute to the introduction of a self-regulation auxiliary system that is developed based on the theories of the positive system and the finite-time stability. Furthermore, the neural network (NN) is utilized to approximate the system uncertainty, and the noncertainty-equivalent structure is adopted to construct the update law of the NN. Then, the influence of the quasi-periodic DoS attack and the appointed ET mechanism on uncertainty approximation are totally isolated. Based on the Lyapunov function method and the invariant set theory, the boundedness of the closed-loop system is theoretically proved. Finally, both the numerical and experimental illustrations are presented to demonstrate the effectiveness and the benefits of the developed control scheme.