EE4-26 Estimation and Fault DetectionLecturer(s): Prof Richard Vinter
The aims are to acquaint students with the need, in various branches of engineering, to estimate the state of a dynamic system from noisy measurements, and also to detect the occurrence of faults and abrupt system changes, and to equip them with some of the principal techniques available for this purpose.
The course enables a student to design a 'Kalman filter' for the least squares estimation of the state of a stochastic system in the linear case, and to design other, suboptimal, filters for use in situations where the either the system or measurement equations are nonlinear . This student will also learn how construct hypotheses tests to decide whether or not system outputs are consistent with the occurrence of a fault.
This course is concerned with the estimation and regulation of linear stochastic systems and the recursive detection of faults and changes. Topics covered include: Stochastic models: power spectra, ARMAX models, state - space representations, covariance equations. Kalman filtering: the Kalman filter, steady-state filters, Bayes confidence intervals, the extended Kalman filter. Fault detection and estimation: Bayes and Neyman-Pearson hypothesis tests, joint estimation and model testing.
One 3-hour exam in April/May
Coursework contribution: 0%
Closed or Open Book (end of year exam): Closed
Oral Exam Required (as final assessment): N/A
Prerequisite: None required
Course Homepage: unavailable