|
|
自动化学报 2005
OnRobustH2Estimation, PP. 1-12 Keywords: Discrete-timesystems,uncertainsystems,robustH2estimation,Riccatiequations,linearmatrixinequalities Abstract: ？Theproblemofstateestimationforuncertainsystemshasattractedarecurringinterestinthepastdecade.Inthispaper,weshallgiveanoverviewonsomeoftherecentdevelopmentintheareabyfocusingontherobustH2(Kalman)filteringofuncertaindiscrete-timesystems.TherobustH2estimationisconcernedwiththedesignofafixedestimatorforafamilyofplantsunderconsiderationsuchthattheestimationerrorcovarianceisofaminimalupperbound.Theuncertaintyunderconsiderationincludesnorm-boundeduncertaintyandpolytopicuncertainty.Inthefinitehorizoncase,weshalldiscussaparameterizeddifferenceRiccatiequationapproachforsystemswithnorm-boundeduncertaintyandpinpointthedifferenceofstateestimationbetweensystemswithoutuncertaintyandthosewithuncertainty.Intheinfinitehorizoncase,weshalldealwithboththenorm-boundedandpolytopicuncertaintiesusingalinearmatrixinequality(LMI)approach.Inparticular,weshalldemonstratehowtheconservatismofdesigncanbeimprovedusingaslackvariabletechnique.Wealsoproposeaniterativealgorithmtorefineadesignedestimator.Anexamplewillbegiventocompareestimatorsdesignedusingvarioustechniques.
|
