In this paper, a method of simultaneous estimation of actuator and sensor faults based on a class of uncertain time-delayed Markovian jump systems is presented. The transition probability matrix (TPM) is assumed to be uncertain. By extending the sensor faults as an auxiliary state, the system is converted into a generalized descriptor system with Markovian jump parameters. Based on the transformed system, a Markovian jump observer is designed to provide the estimates of the states and sensor faults. Meanwhile, a set of adaptive laws are given to adjust the actuator faults. Sufficient conditions of existence of the observer are given in terms of a linear matrix inequality optimization problem. Finally, two numerical examples are simulated to validate the effectiveness of the proposed approach.%针对具有参数不确定和延迟环节的马尔科夫跳变系统,在状态转移概率矩阵(Transition probability matrix, TPM)不确定的情形下,讨论了其执行器和传感器故障同时估计的方法。通过扩展系统状态,将系统转换为一个具有马尔科夫跳变参数的广义描述系统,基于此广义描述系统设计马尔科夫跳变观测器实现对其状态和传感器故障的估计。与此同时,还设计了一组自适应律对执行器故障进行在线调节。通过求解一组线性矩阵不等式最优化问题,得到观测器存在的充分条件。最后,针对两个数值实例,验证了所设计方法的有效性。
展开▼
