Security Analysis and Defense of Multi-Encoding Mechanism Against Eavesdropping Attacks

This paper investigates a defense strategy of remote state estimation against eavesdropping attacks for cyber-physical systems. To prevent interception of transmitted data by an eavesdropper, a random multi-encoding mechanism based on Markov model is proposed, which combines linear transformation and artificial noise. An insecure conditions under which the eavesdropper can deduce the encoding parameters of the multi-encoding mechanism are obtained based on the magnitude of the artificial noise in different eavesdropping scenarios. Furthermore, a method is developed to prevent the eavesdropper from deriving encoding parameters through a novel design of the encoding protocol. It is demonstrated that, even if the eavesdropper obtained some useful transmission data, the security of all the transmitted data can still be guaranteed without knowing the transition probability matrix, which provides a theoretical basis for the design of the multi-encoding mechanism. A simulation example is finally presented to verify the effectiveness and practicality of the proposed method.