NONLINEAR SYSTEM IDENTIFICATION
System identification of nonlinear state space models is a critical component in the development of state estimation and control schemes due to the underlying dependance upon an priorly estimated model. My interest is on the development of robust and efficient methods to perform system identification of generalised nonlinear state space models by performing accurate approximations to this generally intractable problem with the use of Gaussian mixtures.
NONLINEAR STATE ESTIMATION
My research into nonlinear state estimation is focused upon performing approximations to Bayesian filtering and smoothing for nonlinear state space models. These approximations are being formed using Gaussian mixtures in order to allow non Gaussian and potentially multimodal probability densities to be maintained in a computational efficient manner. These approximations can then be used to improve the system identification problem.