A scenario-based approach to parameter estimation in state-space models having quantized output data

DAMIÁN MARELLI; BORIS GODOY; GRAHAM GOODWIN

Atlanta

Conferencia; IEEE Conference on Decision and Control (CDC); 2010

Institute of Electrical and Electronics Engineers (IEEE)

In this paper we describe an algorithm for estimat-ing the parameters of a linear, discrete-time system, in state-space form, having quantized measurements. The estimationis carried out using the maximum likelihood criterion. Thesolution is found using the expectation maximization (EM)algorithm. A technical difficulty in applying this algorithmfor this problem is that the a posteriori probability densityfunction, found in the EM algorithm, is non-Gaussian. To dealwith this issue, we sequentially approximate it using scenarios,i.e., a weighted sum of impulses which are deterministicallycomputed. Numerical experiments show that the proposedapproach leads to a significantly more accurate estimation thanthe one obtained by ignoring the presence of the quantizer andapplying standard estimation methods.