Matrix Riccati Equations in Con- trol and Systems Theory by

Security and reliability are of greatimportance in control system tech-nology. To ensure a reasonable levelof security and reliability, the needfor failure detection (diagnosis)techniques has long been recog-nized. Roughly speaking, a failure isany kind of malfunction of the sys-tem that leads to unacceptable per-formance. Failure detection consistsof techniques for determiningwhether a failure has occurred.These techniques provide themeans for avoiding major break-downs and enable the system opera-tor to take appropriate action.Failure detection is used in high-per-formance systems such as aircraftcontrol systems, power plants, andchemical plants. Not surprisingly,the development of reliable failuredetection techniques has become anintensive field of research. In model-based failure detection,the assumption is that a mathemati-cal model of the behavior of the sys-tem under surveillance, in theabsence of failure, is available. Fail-ure detection then consists of decid-ing whether or not the measuredinputs and outputs are consistentwith this model. In most cases, a real-istic model must account for uncer-tainty and noise, which implies thatthe input-output behavior is not justa single time trajectory but rather aset of time trajectories. This set isoften difficult to characterize, makingthe detection decision difficult.Failure detection has been widelystudied, and the literature includesnumerous books [3], [4], [7]–[9],[15], [20], [21] and survey articles[2], [10]–[14], [16], [19], [23]. In par-ticular, the book [8] gives an up-to-date overview of the subject. Thebook reviewed here presents model-based methodologies for fault detec-tion and fault-tolerant control; someof the methodologies can be found inthe above literature while others arebased on recent research papers.