Covert Attack Detection Based on Hi/H∞ Optimization for Cyber-Physical Systems

Abstract

In this paper, a new detection scheme is proposed to detect covert attack in the framework of Hi/H∞ index optimization for cyber-physical system (CPS) which is modeled as a linear discrete time-varying (LDTV) system. First, a random modulation matrix that the attacker cannot know is inserted into the path of the control variables to destroy the stealthiness of covert attacks. Second, a detection filter is constructed which transforms the detection problem into an H-/H∞ or H∞/H∞ index optimization problem. The optimal solution is obtained by solving the Riccati equation. Third, a decision making mechanism is presented to trigger an alarm and further determine whether the cause of alarm is a covert attack or a fault. Finally, a simulation example is given to illustrate the effectiveness of the proposed method.