Complete stealthiness false data injection attack design in multisensor systems with attack resource constraints

Abstract

AbstractThis article is concerned with the attack design problem of multisensor cyber‐physical systems. Compared with the existing results developed for the single‐sensor system, the attacker is limited to modify partial sensors. To avoid generating alarm for the given detect mechanism, the correlation influence between the attacked and normal sensors is eliminated in the designed stealthiness condition. Then the trace of error covariance is used to measure the attack effect and the coupling information between sensors is also involved. By using the design freedom of stealthiness condition, the detailed attack parameters design and attacked channels selection strategy are derived to achieve the optimal attack performance via a two‐stage method including backstepping and eigenvalue sorting techniques. Finally, numerical simulation examples are provided to verify the effectiveness and advantages of the proposed attack method.