Abstract:
In this presentation, we will discuss recent advance on identification algorithms for finiteimpulse response systems under quantized output observationsand general quantized inputs. While asymptotically efficientalgorithms for quantized identification under full-rank periodic inputsare available, their counterpart under general inputs hasencountered technical difficulties and evadedsatisfactory resolutions for a decade. Under quantized inputs, our recent results have resolved this issue with constructive solutions.A two-step algorithm is developed, which demonstrates desiredconvergence properties includingstrong convergence, mean-square convergence, convergence rates,asymptotic normality, and asymptotical efficiency in terms of the Cramer-Rao lower bound. Some essential conditions on inputexcitation are derived that ensure identifiability and convergence.It is shown that by a suitable selection of the algorithm's weighting matrix,the estimates become asymptotically efficient. The strong andmean-square convergence rates areobtained.Optimal input design is given.Also the joint identification of noise distribution functionsand system parameters is investigated.Numerical examples are included to illustrate the main results.