Efficient Computation of the Continuous-Discrete Extended Kalman Filter Sensitivities Applied to Maximum Likelihood Estimation
- 1 December 2019
- conference paper
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 6983-6988
Abstract
In this paper, we present and compare different methods for computing the likelihood function and its gradient. We consider nonlinear continuous-discrete models described by a system of stochastic differential equations (SDEs) with discrete-time measurements. The problem of maximum likelihood estimation (MLE) is formulated as a nonlinear program (NLP) and it is solved numerically using a gradient-based single shooting algorithm. The estimates of the mean and its covariance are computed using a continuous-discrete extended Kalman filter (CDEKF). We derive analytical expressions for the gradient of the likelihood function. We discuss some aspects of the implementation of MLE for non-stiff systems. In particular, we present an efficient way of computing the state covariance matrix and its gradient using explicit Runge-Kutta schemes. We verify our implementation using a numerical example related to type 1 diabetes and demonstrate how to apply it for nonlinear parameter estimation.Keywords
This publication has 27 references indexed in Scilit:
- How To NOT Make the Extended Kalman Filter FailIndustrial & Engineering Chemistry Research, 2013
- Maximum likelihood estimation of inflation factors on error covariance matrices for ensemble Kalman filter assimilationQuarterly Journal of the Royal Meteorological Society, 2011
- A particle filter approach to identification of nonlinear processes under missing observationsThe Canadian Journal of Chemical Engineering, 2008
- Parameter estimation in stochastic grey-box modelsAutomatica, 2004
- A direct derivation of the exact Fisher information matrix of Gaussian vector state space modelsLinear Algebra and its Applications, 2000
- Analytic derivatives for estimation of linear dynamic modelsComputers & Mathematics with Applications, 1989
- Computation of the exact information matrix of Gaussian time series with stationary random componentsIEEE Transactions on Acoustics, Speech, and Signal Processing, 1986
- Numerical Treatment of Inverse Problems in Chemical Reaction Kinetics1981
- Maximum likelihood and prediction error methodsAutomatica, 1980
- A family of embedded Runge-Kutta formulaeJournal of Computational and Applied Mathematics, 1980