Transfer-function models of sampled systems

Finite-length Padé approximants to sampling delays enable sampled systems to be represented by continuous models. Accuracy depends on the length of the rational Padé approximants, since they determine the complexity of the models. System-reduction techniques provide a compromise between accuracy and complexity. Low-order Padé-based models provide conveniently simple indicators of distortion induced by digital discretisation of a continuous system in terms of system-parameter perturbations and dead time. System-parameter and time-shift compensations can greatly reduce the distortion.