中文摘要：

汽车稳定控制系统具有明显的非线性特性,对其传感器与执行器的故障诊断具有一定难度。从汽车八自由度模型入手,简化模型并建立虚拟输入。针对虚拟输入模型,将虚拟执行器故障作为增广状态向量的一部分,构造一个增广系统。利用李雅普诺夫理论设计观测器,获得原系统状态与虚拟执行器故障的渐近估计,并分析了增广观测器的稳定性与收敛性。通过LMI技术实现线性矩阵不等式求解,完成观测器设计。选择一个新的状态变量作为输出信号滤波器,将传感器故障转化为执行器故障,建立二增广系统,直接利用执行器故障诊断方法实现传感器故障诊断。通过Simulink,验证了此方法的可行性。与其他方法相比,该方法不仅能够同时得到虚拟执行器与传感器故障信号,为系统提供警示,而且为后续的容错控制设计提供可靠数据。

英文摘要：

Vehicle stability control system features obvious nonlinear characteristics.It is difficult for fault diagnosis of its sensor and actuator.The eight degree of freedom model of vehicle is simplified and the virtual input is established.Aiming at the virtual input model,the actuator fault is considered as part of augmented state vector,and an augmented system is constructed.The observer is designed by using Lyapunov theory; the asymptotic estimation of the state of original system and the virtual actuator fault is obtained,and the stability and convergence of the augmented observer is analyzed.Through LMI technology,the solution of inequalities of linear matrix is realized,the design of observer is completed.By adopting simple filter,a new state vector is selected as low-pass 1st order filter; the sensor fault is transformed into actuator fault; the 2nd augmented system is constructed,then the fault estimation of sensor is implemented by directly using actuator fault estimation.The feasibility of this method is verified by Simulink.Comparing with other methods,this method can obtain the fault signals of virtual actuator and sensor simultaneously for warning the system,and provide reliable data for subsequent design of fault tolerant control.