Recursive Privacy-Preserving Estimation Over Markov Fading Channels

TL;DR

This paper proposes a recursive, privacy-preserving estimation method for systems over Markov fading channels, ensuring secure state estimation against eavesdroppers through a novel encoding and filtering strategy, validated on a three-tank system.

Contribution

It introduces a new co-design approach combining privacy mechanisms with recursive filtering for secure estimation over Markov fading channels.

Findings

The proposed method maintains bounded estimation error for legitimate users.

The encoding scheme effectively prevents eavesdropper's estimation accuracy.

Simulation confirms the method's effectiveness on a three-tank system.

Abstract

In industrial applications, the presence of moving machinery, vehicles, and personnel, contributes to the dynamic nature of the wireless channel. This time variability induces channel fading, which can be effectively modeled using a Markov fading channel (MFC). In this paper, we investigate the problem of secure state estimation for systems that communicate over a MFC in the presence of an eavesdropper. The objective is to enable a remote authorized user to accurately estimate the states of a dynamic system, while considering the potential interception of the sensor's packet through a wiretap channel. To prevent information leakage, a novel co-design strategy is established, which combines a privacy-preserving mechanism with a state estimator. To implement our encoding scheme, a nonlinear mapping of the innovation is introduced based on the weighted reconstructed innovation previously received by the legitimate user. Corresponding to this encoding scheme, we design a recursive privacy-preserving filtering algorithm to achieve accurate estimation. The boundedness of estimation error dynamics at the legitimate user's side is discussed and the divergence of the eavesdropper's estimation error is analyzed, which demonstrates the effectiveness of our co-design strategy in ensuring secrecy. Furthermore, a simulation example of a three-tank system is provided to demonstrate the effectiveness and feasibility of our privacy-preserving estimation method.