Optimal Fusion Deception Attacks Design and Game Analysis in Multi-Sensor Systems

In this paper, the problems of designing fusion deception attacks against multi-sensor systems are investigated. Some new fusion deception attack models have been designed. The trace of the estimation error is used to evaluate the attack performance and the Kullback-Leibler (K-L) divergence is utilized to evaluate the stealthiness. The benchmarking is to compare the estimation errors under the same stealthiness constraint. It has been verified that the proposed attack schemes perform better than the existing ones. Besides, the consideration of correlation help to improve the attack performance as well. A game framework was also established to study the interaction between the attacker and the defender. Based on the game analysis, optimal attack scheme and defense policy are simultaneously derived to achieve a Nash equilibrium. Finally, simulation results are provided to illustrate the theoretical results.