A Bayesian Stackelberg Game Approach to Remote State Estimation Under SINR-Based DoS Attacks with Incomplete...

Area of Science:

• Cyber-Physical Security
• Game Theory
• Information Theory

Background:

• Denial-of-Service (DoS) attacks pose significant threats to cyber-physical systems, particularly impacting remote state estimation.
• Limited energy resources necessitate efficient transmission and interference strategies for sensors.
• Incomplete information about attacker channel gains complicates defense mechanisms.

Purpose of the Study:

• To investigate Stackelberg game strategies for a sensor and an attacker in remote state estimation under signal-to-interference-plus-noise ratio (SINR)-based DoS attacks.
• To develop a method for balancing sensor estimation performance and energy consumption.
• To address scenarios with incomplete information regarding the attacker's channel gain.

Main Methods:

• Formulation of the scheduling problem as a Markov decision process over an infinite-time horizon.
• Construction of a Bayesian Stackelberg game (BSG) incorporating probabilistic channel gain information.
• Application of the best-response definition to solve the BSG and prove the existence of a Stackelberg equilibrium.
• Utilization of a Stackelberg Q-learning algorithm to derive optimal strategies for both sensor and attacker.

Main Results:

• The proposed Bayesian Stackelberg game effectively models the strategic interactions between the sensor and the attacker.
• The Stackelberg Q-learning algorithm successfully identifies optimal transmission and interference strategies.
• Numerical results validate the game method's effectiveness, even when the sensor lacks precise knowledge of the attacker's channel gain.

Conclusions:

• The developed game-theoretic approach provides an effective framework for defending remote state estimation against DoS attacks under energy constraints and incomplete information.
• The findings highlight the importance of adaptive strategies in cyber-physical security.
• The study demonstrates a viable solution for enhancing system resilience in adversarial environments.