作者: Giulio Bottegal , Riccardo S. Risuleo , Mohsen Zamani , Brett Ninness , Håkan Hjalmarsson
DOI: 10.1016/J.IFACOL.2017.08.634
关键词: Expectation–maximization algorithm 、 Likelihood function 、 Estimation theory 、 Algorithm 、 Maximum likelihood sequence estimation 、 Errors-in-variables models 、 Restricted maximum likelihood 、 Marginal likelihood 、 M-estimator 、 Computer science
摘要: Abstract In this paper, we revisit maximum likelihood methods for identification of errors-in-variables systems. We assume that the system admits a parametric description, and input is stochastic ARMA process. The cost function associated with criterion minimized by introducing new iterative solution scheme based on expectation-maximization method, which proves fast easily implementable. Numerical simulations show effectiveness proposed method.