Prescribed performance-based resilient model-free adaptive control for CPSs against aperiodic DoS attacks

This paper presents a prescribed performance-based model-free adaptive resilient control (MFARC) algorithm to address the realization of trajectory tracking for nonlinear cyber-physical systems (CPSs) under aperiodic denial-of-service (DoS) attacks. Firstly, an equivalent linear model is constructed by utilizing the dynamic linearization technique, and a novel output transformation approach is applied to convert the constrained system to an unconstrained one for the MFARC framework. Then, an attack compensation mechanism is presented to estimate the unavailable output signal when attacks are active, such that the negative impact of aperiodic DoS attacks can be compensated. Based on the unconstrained model and the compensation mechanism, a prescribed performance-based MFARC strategy is constructed to steer the tracking error to the predetermined neighborhood around the origin in the presence of DoS attacks. A major improvement of our result over previous studies is the introduction of prescribed performance control (PPC) into MFARC while reasonably restricting the transient and steady-state performance of the tracking error even when DoS attacks occur. Additionally, we utilize the linear matrix inequality (LMI) toolbox to obtain better tracking performance by suitably adjusting the time-varying parameters, which is superior in practical applications. Ultimately, our results indicate that the tracking error consistently remains within a predefined bound.