Connections between incremental and continuous-time EM algorithm for state space identification

Abstract The EM algorithm has been successfully applied to obtain maximum likelihood estimates for state-space models. The usual formulation of the algorithm is based on a shift operator model for the discrete-time (or sampled-data) system. More recently, it has been shown that an equivalent formulation of the algorithm in terms of incremental discrete-time models shows better numerical properties, in particular, for fast sampling rates. In this paper we explore the correspondence between the parameter estimates given by the EM algorithm applied to incremental models and those corresponding to a purely continuous-time formulation.

[1]  P. Fearnhead,et al.  Exact and computationally efficient likelihood‐based estimation for discretely observed diffusion processes (with discussion) , 2006 .

[2]  G. Goodwin,et al.  Insights into the zero dynamics of sampled-data models for linear and nonlinear stochastic systems , 2007, 2007 European Control Conference (ECC).

[3]  G. Goodwin,et al.  Frequency domain identification of MIMO state space models using the EM algorithm , 2007, 2007 European Control Conference (ECC).

[4]  D. Rubin,et al.  Maximum likelihood from incomplete data via the EM - algorithm plus discussions on the paper , 1977 .

[5]  Graham C. Goodwin,et al.  On sampled-data models for nonlinear systems , 2005, IEEE Transactions on Automatic Control.

[6]  Graham C. Goodwin,et al.  Sampling and sampled-data models , 2010, Proceedings of the 2010 American Control Conference.

[7]  N. Shephard,et al.  Likelihood INference for Discretely Observed Non-linear Diffusions , 2001 .

[8]  Yacine Aït-Sahalia.  Maximum Likelihood Estimation of Discretely Sampled Diffusions: A Closed‐form Approximation Approach , 2002 .

[9]  Lennart Ljung,et al.  Issues in sampling and estimating continuous-time models with stochastic disturbances , 2010, Autom..

[10]  A. Dembo,et al.  Parameter estimation of partially observed continuous time stochastic processes via the EM algorithm , 1992 .

[11]  G. Goodwin,et al.  SAMPLED-DATA MODELS FOR STOCHASTIC NONLINEAR SYSTEMS , 2006 .

[12]  Brett Ninness,et al.  Robust maximum-likelihood estimation of multivariable dynamic systems , 2005, Autom..

[13]  H. Sørensen.  Parametric Inference for Diffusion Processes Observed at Discrete Points in Time: a Survey , 2004 .

[14]  G. Goodwin,et al.  Identification of continuous-time state-space models from non-uniform fast-sampled data , 2011 .

[15]  Adrian Wills,et al.  Issues in sampling and estimating continuous-time models with stochastic disturbances , 2008 .

[16]  Hermann Singer,et al.  CONTINUOUS-TIME DYNAMICAL SYSTEMS WITH SAMPLED DATA, ERRORS OF MEASUREMENT AND UNOBSERVED COMPONENTS , 1993 .

[17]  Brett Ninness,et al.  Maximum Likelihood estimation of state space models from frequency domain data , 2009, 2007 European Control Conference (ECC).