Parametric identification of a linear time invariant model for a subglottal system

Models of the human body are key in bio-engineering and medical applications. This study presents the identification, in time and frequency domains, of linear time invariant models of the human subglottal system, for the clinical assessment of vocal function. For time domain identification, the input-output data corresponds to the glottal volume velocity and the acceleration registered by a sensor on the neck skin of the patient. For frequency domain identification, the frequency response of the subglottal tract module is used. Maximum likelihood and prediction error methods are applied. Additionally, the Akaike and Bayes Information Criteria are used to select the models order.