Expectation maximization estimation algorithm for Hammerstein models with non-Gaussian noise and random time delay from dual-rate sampled-data

This paper considers the robust identification for dual-rate input nonlinear equation-error systems with outliers and random time delay. To suppress the negative influence caused by the outliers to the accuracy of identification, the distribution of the noise is represented by a t-distribution rather than a Gaussian distribution. A random time delay is considered in the dual-rate input nonlinear systems. By treating the unknown time delay as the latent variable, the expectation maximization algorithm is derived for identifying the systems. Two numerical simulation examples demonstrate that the proposed algorithm can generate accurate identification results when the measurements are contaminated by the outliers.