Slide 1OutlineIntroObjectivesSystem ModelSlide 6Slide 7Recap…Test δaSlide 10Test δbSlide 12Recap…Asymptotic OptimalitySlide 15Higher-order ApproximationSlide 17ConclusionReferencesSlide 20Slide 21Higher-order ApproximationSlide 23Expected OvershotSlide 25Expected r_maxD_iAsymmetric of MSPR TestsSlide 29Slide 30Multihypothesis Sequential Probability Ratio TestsPresented by: Yi (Max) HuangAdvisor: Prof. Zhu HanWireless Network, Signal Processing & Security LabUniversity of Houston, USA2009 11 12Published by V. P. Dragalin, A. G. Tartakovsky, and V. V. VeeravalliPart I: Asymptotic OptimalityPart II: Accurate Asymptotic Expansions for Expected Sample SizeOutlineIntroductionObjectivesSystem ModelMSPR TestsTest δa – Bayesian Optimality Test δb – LLR testAsymptotic Optimality of MSPR TestsThe case of i.i.d. ObservationsHigher-order Approximation for accurate resultsConclusionsIntroMSPRT – Multiple hypothesis + SPRTMany applications need M≥2 hypothesisMultiple-resolution radar systemsCDMA systemsThe quasi-Bayesian MSPRT with expected sample size in asymptotical i.i.d case was established by V. Veeravalli.Need…make accurate decision by shortest observation time for MSPRTObjectivesContinue investigate the asymptotic behaviors for two MSPRTsTest δa quasi-Bayesian optimalityTest δb log-likelihood-ratio (LLR) testGoal:Asymptotically optimal to any positive moment of stopping time distribution.minimize average observation time and generalize the results for most environments.System ModelThe sequential testing of M hypotheses ( ).. is a sequential hypothesis test. d is “terminal decision function”; d=i, Hi occurs is the prior distribution hypotheses.loss function: In case of zero-one : is the conditional error probability.Ri is the risk function:when and =Pr[accepting Hi incorrectly] The class of test: The predefined value:Z is the log-likelihood function and ratio (up to given time n) is the positive threshold and is defined by is the expected stopping time (average stopping time)Recap…IntroductionObjectivesSystem ModelMSPR TestsTest δa – Bayesian Optimality Test δb – LLR testAsymptotic Optimality of MSPR TestsThe case of i.i.d. ObservationsHigher-order Approximation for accurate resultsConclusionsTest δa and is applied by a Bayesian framework.If , thenIf , thenIn the special case of zero-one loss: Remark: stop as soon as the largest posterior probability exceeds a threshold, A.Test δb and is corresponded to LLR testvi is the accepting time for HiIn the special case of zero-one loss:If , thenRecap…IntroductionObjectivesSystem ModelMSPR TestsTest δa – Bayesian Optimality Test δb – LLR testAsymptotic Optimality of Ei[T] in MSPRTThe case of i.i.d. ObservationsHigher-order Approximation for accurate resultsConclusionsAsymptotic Optimality is Kullback-Leibler (KL) information distance is minimal distance between Hi and othersConstraints: Dij must be positive, finiteRi to 0Asymptotic lower bounds for any positive moments of stopping time:Now, the minimal expected stopping time in asymptotic optimality:Higher-order ApproximationIn order to has accurate results, the higher order approx. expected stopping time Is estimated:Using non-linear renewal theory, & are now in form of random walk crossing a constant boundary + the non-linear term “slowly changing”The result of “slowly changing” is that limiting overshot (xi) of random walk over a fixed threshold are unchangedRedefined to When “r=M-1” and “\i” exclusion of “i” from the sethr,i is expected value of max r zero-mean normal r.v.The expected stopping time in higher-order approx. for Test δb and Test δa :ConclusionThe proposed MSPRT are asymptotically optimal under fairly general conditions in discrete or continuous time, stochastic models, and, …etc.Asymptotically optimal minimize any positive moment of the stopping time (average observation time) in both generalized approx. by “risk” go to zero, and higher-order approximations , up to a vanishing term by non-linear renewal theory.ReferencesV. P. Dragalin, A. G. Tartakovsky, and V. V. Veeravalli, “Multihypothesis Sequential Probability Ratio Tests – Part I: Asymptotic Optimality”, IEEE TRANSACTIONS ON INFORMATION THEORY, VOL. 45, NO. 7, NOV. 1999V. P. Dragalin, A. G. Tartakovsky, and V. V. Veeravalli, “Multihypothesis Sequential Probability Ratio Tests – Part II: Accurate Asymptotic Expansions for the Expected Sample Size”, IEEE TRANSACTIONS ON INFORMATION THEORY, VOL. 45, NO. 7, NOV. 1999Higher-order ApproximationIn order to has accurate results, the higher order apprex. expect stopping time Is estimated:Nonlinear to randomExpected OvershotExpected r_maxD_irelax the condition , and rewrite it to:Asymmetric of MSPR TestsSymmetric case:guarantees -to-zero rate keeping up ‘s rate,
View Full Document