Abstract

This paper addresses the challenging problem of strictly enforcing continuous-time path constraints for nonlinear systems over the entire time horizon under aperiodic denial-of-service (DoS) attacks. We propose a novel resilient dual-mode (comprising outer and inner modes) event-triggered model predictive control (MPC) framework. In the outer mode, a resilient event-triggered scheme is designed to adaptively schedule controller updates by integrating information on the prediction horizon, DoS attack patterns, real-time variations in path constraints, and mode transitions of the system state. Subsequently, a path-constrained event-triggered MPC algorithm is developed through iterative application of the right-hand restriction method for semi-infinite programming, which guarantees constraint satisfaction while steering the system state into the inner mode. In the inner mode, an additional resilient event-triggered scheme driven by real-time state variations is introduced. A state-feedback controller and the closed-loop system model are then established by exploiting the inclusion relationship between triggering intervals and attack durations. Furthermore, a unified resilient dual-mode event-triggered path-constrained MPC algorithm is formulated, alongside sufficient conditions to guarantee closed-loop stability. Finally, numerical simulations demonstrate the effectiveness of the proposed approach.

DOI: https://doi.org/10.1109/JIOT.2026.3742508