An optimized lumped model for tracheoesophageal speakers

Abstract

Since the advent of the first lumped model of the phonatory system by Flanagan and Landgraf (1968), many improvements have been made in order to capture more realistic features of the human phonation in glottal speakers. Although lumped models are quite limited when it comes to representing realistic structural and fluid dynamic aspects, they are very important to understand fundamental features of human phonation. Nevertheless, only a few lumped models have been proposed in order to investigate the dynamics of the pharyngoesophageal segment in tracheoesophageal speakers. This work presents some improvements on a lumped model of the pharyngoesophageal segment, firstly proposed by Schwarz et al. (2011). The improvements have been achieved by conducting an optimization process that involves the glottal volume flow as the optimization function, using as reference the experimental data obtained from tracheoesophageal speakers. The results suggest that the lumped model of the esophageal region can be extended to capture important features, such as the pressure threshold for the onset of self-sustained oscillation.