Random Reverse Multi-Encoding Mechanism on Remote State Estimation

Abstract

This letter studies a reverse multi-encoding mechanism based on Markov model to resist eavesdropping attacks for remote state estimation in cyber-physical systems. The innovation is transmitted between the sensor and the remote state estimation over insecure and unreliable networks. An eavesdropper collects the transmitted innovation data from the eavesdropping network. To maintain the confidentiality of the transmitted innovation, a reverse multi-encoding mechanism based on Markov model is proposed to resist eavesdropping attacks, where the Markov model is pre-set at the sensor and remote state estimation. Based on the open-loop performance and the Markov model, a lower bound on the transmission probability of the encoded innovation is given, which can effectively ensure that the estimation error covariance for the eavesdropper is higher than that for the legitimate user. An uncrewed aerial vehicle is used to illustrate the effectiveness and practicality of the proposed reverse multi-encoding mechanism based on the Markov model.