An event-triggered reliable cloud control scheme based on ADP and integral sliding mode

Abstract

This paper investigates event-triggered (ET) reliable control problems for cloud control systems under actuator faults and data injection attacks via the adaptive dynamic programming (ADP) and integral sliding mode (ISM). A mist-fog-regional cloud control architecture is first given, which can improve computing efficiency of the cloud platform. In this architecture, a fog-based fault parameter estimation method is proposed with the aid of neural networks. It is driven by the feedback of fault parameter estimation errors, so as to achieve more accurate estimations of fault parameters. A double ET reliable cloud control scheme is further presented. It is composed of an ISM-based and an ADP-based regional cloud controllers. As a result, it not only saves communication resources, but also eliminates the influence of the attacks and matched uncertainties, as well as ensures the stability of the equivalent sliding-mode dynamics with optimal performance. Finally, the effectiveness of the proposed method is verified by the simulation results.