Physiological modeling of speech production: methods for modeling soft-tissue articulators.

As the initial components of a comprehensive physiological model of speech production, methods have been developed for modeling soft tissue structures, such as the tongue and lips, as continua. This approach provides a foundation for applying finite element methods to simulate these structures in a biomechanical model of speech production. Movements and deformations of the structures can then be computed as the solutions of a nonlinear second-order constraint system of ordinary differential equations, which is obtained from a finite element approximation of an energy rate equation. The muscle fibers in the soft tissue structures are represented as fields that specify the directions in which active and passive tensile stress is produced. The elastic behavior of the passive components is modeled using an isotropic exponential constitutive strain energy function, and the viscous stress components, by using linear viscosity. Incompressibility is maintained elementwise in the simulation using a simultaneously solved system for the computation of Lagrange multipliers. As first step towards modeling a vocal tract, a preliminary model of a tongue with eight muscles has been completed. Simulation results demonstrate the validity of the method and they support the feasibility of a physiologically based model of speech production.