H-infinity estimation, some old and new results
Prof. Uri Shaked (Tel Aviv and CUED)
H-infinity optimal estimation has been developed during the last decade.
Its advantages stem from its low
sensitivity to the exact knowledge of the system model and the
statistics of the exogenous signals. It has soon
been noticed that this type of estimation can readily cope with a
significant uncertainty in the parameters of the
process to be estimated, and optimal design techniques have been
suggested that guarantee a required estimation
level in spite of a polytopic type uncertainty in the linear case or
norm bounded uncertainties in the linear and the nonlinear cases.
A short survey of H-infinity estimation will be given with a special
emphasis on the latest results on robust H-infinity filtering.
Back to
Control Seminars Page.