A novel UIO-based approach for fault detection and isolation in finite frequency domain

This paper proposes a novel unknown input observer (UIO) design approach to detect and isolate the actuator faults in finite frequency domain. Instead of completely rejecting the disturbance or noise of traditional UIOs, we exclude certain controlled inputs from the residuals. Meanwhile, knowledge on the disturbance/noise, such as frequency and statistical characteristics, is embedded into the design to enhance the performance of fault detection and isolation. One significant advantage of our approach lies in the number of UIOs which is dramatically reduced when considering multiple faults. All the design conditions are formulated in terms of LMI or BMI problems, which enable to consider other design specifications in addition to the robustness and sensitivity requirements.