Fault-Tolerant Observer and Trajectory Tracker for Sampled-Data Systems: Digital Redesign Approach

• Main Authors: Chao-LungWei, 魏兆隆
• Other Authors: Sheng-Hong Tsai
• Format: Others
• Language: en_US
• Published: 2010
• Online Access: http://ndltd.ncl.edu.tw/handle/70634486547627477370

博士 === 國立成功大學 === 電機工程學系碩博士班 === 98 === Digitally redesigned adaptive observer, fault-tolerant observer, and trajectory tracker for some classes of sampled-data systems with actuator and state faults have been proposed in this dissertation to obtain an efficient fault detection and performance recovery. Here, it can be stated in following aspects: the evolutionary programming-based digitally redesigned adaptive observer, fault estimator, and trajectory tracker for the actuator fault detection and performance recovery via the input-compensation scheme have been established; the prediction-based digitally redesigned adaptive observer, which uses the currently measured output to estimate the current state for sampled-data linear time-varying system, is proposed to achieve a high-performance recovery for the faulted system with actuator and state faults. For achieving this goal, a residual generation scheme and a mechanism for auto-tuning switched gain to fit either the actuator or state fault scenario are proposed; the predictive Kalman filter-based adaptive observer and trajectory tracker for sampled-data nonlinear time-varying system have been proposed to automatically detect the types of faults (i.e., state fault or actuator fault), the amount of unanticipated fault in terms of decay factors, and compensate for it. The utilization of optimal linearization method accomplishes the proposed methodology for actuator fault detection and performance recovery for sampled-data nonlinear time-varying systems. The effectiveness of the digitally redesigned adaptive observer and trajectory tracker for performance recovery of the faulted systems is also demonstrated by illustrative examples.