Attack-Aware Event-Triggered Impulsive Control for Switched Nonlinear Systems via Multi-Diffeomorphism

Abstract

This paper develops a unified hybrid control architecture for switched nonlinear systems under aperiodic denial-of-service (DoS) attacks with non-minimum phase internal dynamics. The methodology relies on three technical pillars: multi-diffeomorphism input-output linearization, a mode-dependent event-triggered impulsive mechanism, and an attack-aware switching rule. The latter maximizes a resilience index during DoS-active intervals, while the event-triggered mechanism reduces communication updates, exploits mode information, and enables rapid correction after attacks. We derive constructive sufficient conditions based on AED-ADT and DoS parameters that guarantee ISS with an explicit exponential decay rate, and establish Zeno-free operation via a uniform positive lower bound on inter-event intervals. We also address the case where a DoS attack coincides with a switching instant and show that the architecture remains applicable. Validation via simulations on a nonholonomic mobile robot and a two-link robotic manipulator confirms superior convergence, resource efficiency, and robustness over baselines.