虚假数据注入攻击下线性参数时变多智能体系统H_(-)/H_(∞)攻击检测

H_(-)/H_(∞) cyber-attack detection for linear parameter time-varying multi-agent systems under false data injection attack

针对线性参数时变多智能体系统,解决了带有数据包丢失的虚假数据注入攻击下的攻击检测问题。构建了H_(-)/H_(∞)性能指标去设计虚假数据注入攻击检测模块,H_(-)性能提高攻击检测器的检测精度;H_(∞)性能增强检测器模块对外部干扰的鲁棒性。网络丢包建模为伯努利二值分布。基于随机稳定性理论,将线性矩阵不等式的可行性问题转化为多目标优化问题进行求解,获得最终的攻击检测器增益。数值仿真结果验证了设计的攻击检测滤波器有效性。

In this paper,we focus on the cyber-attack detection problem in false data injection attacks with packet dropouts scenarios for linear parameter time-varying multi-agent systems.The H_(-)/H_(∞)performance indexes are used to design the false data injection attack detection filter.The H_(-)performance index is used to improve the detection accuracy of the attack detector and the H_(∞)performance index is used to enhance the robustness of the detector module,respectively.Modelling the packet dropout phenomenon as Bernoulli process.According to stochastic stability theory,the feasibility problem of linear matrix inequality is transformed into a multi-objective ultimate problem for solving the optimal detector filter gains.The effectiveness of the designed attack detection filter is verified by numerical simulation.