Abstract: The problem of robust fault detection is investigated for networked uncertain systems with multiple packet dropouts.The uncertain plant is represented by polytopic discrete-time model，and stochastic variables with general probability distributions are utilized to characterize the multiple packet-dropout phenomenon.The conditions for the existence of the robust fault detection filter are proposed by using Lyapunov stability theory of stochastic systems and a parameter-dependent approach.The robust fault detection filter makes the error between the residual signal and the weighed fault signal as small as possible，and its gains can be obtained via the feasible solutions to a set of linear matrix inequalities.Simulation examples illustrate the effectiveness of the proposed method.

Key words: fault detection, networked systems, multiple packet dropouts

摘要： 研究一类存在多个数据包丢失的网络化不确定系统的鲁棒故障检测问题。不确定对象为凸多面体离散时间模型。多个数据丢失现象由具有一般概率密度的随机变量描述。利用随机系统的Lyapunov稳定性理论和参数依赖方法给出了使残差信号与加权故障信号之间的误差尽可能小的鲁棒故障检测滤波器的存在性条件。借助于线性矩阵不等式组的可行解得到了故障检测滤波器增益。仿真例子表明了所给方法的有效性。

关键词: 故障检测, 网络化系统, 多个数据包丢失