A robust and efficient system identification method for a state-space model with heavy-tailed process and measurement noises

In the paper, a robust and efficient system identification method is proposed for a state-space model with heavy-tailed process and measurement noises by using the maximum likelihood criterion. An expectation maximization algorithm for a state-space model with heavy-tailed process and measurement noises is derived by treating auxiliary random variables as missing data, based on which a new nonlinear system identification method is proposed. Noise parameter estimations are updated analytically and model parameter estimations are updated approximately based on the Newton method. The effectiveness of the proposed method is illustrated in a numerical example concerning a univariate non-stationary growth model.