Robustness and failure-tolerance are complementary characteristics that work to achieve satisfactory system performance under both normal and abnormal operating conditions. Because robustness is itself a statistical concept, a probabilistic, numerical framework using Monte Carlo evaluation and genetic algorithms for analysis and design of control systems is pursued. The design approach is equally applicable to both linear and nonlinear systems, and it is not limited in either the number of uncertain parameters or the nature of the uncertainty. Failure-tolerant control is presented as a sequence in which mis-behaving sensors and actuators are detected and identified, and control logic is reconfigured to accommodate the failures. Notions of intelligent control are discussed, and recent applications of a "fuzzy" parity-space approach to failure detection and identification are presented.
(Robert Stengel is a Professor of Mechanical and Aerospace Engineering at Princeton University, a Fellow of the IEEE and the AIAA, a graduate of MIT and Princeton, former associate editor-at-large of the IEEE Transactions on Automatic Control, and North American Editor of the Cambridge University Press Aerospace Series. He is the author of Optimal Control and Estimation (Dover, 1994), and he is writing a book on aircraft flight dynamics and control.)