Problems in Diagnosis and Fault-Tolerant Control

In this talk, we discuss two problems in model-based diagnosis and fault-tolerant control: how to perform fault diagnosis when the system model is incomplete, and how to achieve fault tolerance in the presence of sensor and actuator failures. For the first problem, we consider the design of a fault diagnoser based on a possibly incomplete nominal model of a discrete event system. Standard diagnosis algorithms will halt when inconsistencies arise between the observed behaviour and the predicted behaviour from the nominal model. We solve this problem by introducing the learning diagnoser which makes hypotheses about the missing information as well as diagnose faults. We also discuss under what conditions the learning diagnoser can learn the missing model information, and can diagnose faults after they occur. For the second problem, we consider decentralized control of a large flexible space structure in which sensors and actuators may fail. When failures are known, we design a simple control reconfiguration strategy which ensures that reference tracking and disturbance rejection are achieved for the healthy parts of the system. When failures are not known, we use a passive decentralized control scheme to stabilize the system and to extract fault information. Finally, we combine the passive controller, fault information extraction, and control reconfiguration to produce an overall fault-tolerant control design.
Raymond H. Kwong received the S.B., S.M. and Ph.D. degrees in Electrical Engineering from the Massachusetts Institute of Technology, Cambridge, in 1971, 1972, and 1975, respectively. From 1975 to 1977, he was a visiting Assistant Professor of Electrical Engineering at McGill University and a Research Associate at the Centre de Recherches Mathematiques, Universite de Montreal, Montreal, Canada. Since 1977, he has been with the Edward S. Rogers Sr. Department of Electrical and Computer Engineering at the University of Toronto, where he is a Professor. His current research interests include estimation and stochastic control, adaptive signal processing and control, fault diagnosis and fault-tolerant control, discrete event systems, and multimedia security.