Remote State Estimation with Quantized Communication Against Markov Tampering Attack

This paper investigates the security problem of remote state estimation with communication constraints. In a wireless network where only binary-quantized measurements can be transmitted, we model the adversary's stochastic tampering attack as a Markov process and derive the minimum mean-square error state estimator for this scenario. From the attacker's viewpoint, we design an optimal tampering strategy that maximizes the estimation error at the remote center. From the defender's standpoint, when the attack strategy is known, we propose a compensation algorithm that recovers the estimator's performance to its unattacked level; when the attack strategy is unknown, we develop a joint update mechanism that alternately identifies the true attack strategy and refines the state estimate, and we obtain optimal defense parameters by solving for the extremum of a loss function that balances these two objectives. Finally, we validate relevant theoretical results through numerical simulations.