Quantitative analysis of masticatory performance in vertebrates

To quantitatively measure mechanical performance signals such as forces and motions and mechanical breakdown of food during mastication, it is imperative to accurately reproduce the mastication motion. Reproduction of the mastication motion of a vertebrate with a robotic device will allow us to estimate muscle and bite forces required for different animals while chewing/biting different regimen and relate them to masticatory muscle recruitment patterns and would be used to quantitatively evaluate the dynamic breakdown of foods during chewing, which is vital information required in the development of new pet foods. We also examine the use of a robotic solution where a generic parallel manipulator with six degrees of freedom (Stewart platform) was modeled and simulated using virtual prototyping tools to reproduce the 3D mandible trajectory. To this end, a high fidelity (speed/resolution) motion capture system was used for capture the 3D mastication motion of different vertebrates. 3D laser scanning technology and image processing techniques were used to obtain CAD model of a skull and mandible of a bulldog which was then rapidly prototyped and casted to create a dentition. Architectural parameters of muscle for a human jaw were obtained from Koolstra et al. and for a bulldog jaw by conducting dissection of masticatory muscles. A musculoskeletal model of the vertebrate jaw was created in AnyBody Modeling System to measure the forces acting in the masticatory muscles and temporomandibular joints. We formulate and verify the forward dynamics of the Stewart platform using three methods: (1) S-Function in Simulink (2) DynaFlexPro model (3) Visual Nastran Plant. A combination of Newton-Euler and Lagrangian method was used to formulate the inverse dynamics of a 6 DOF parallel manipulator. Feedback linearization was implemented in Matlab/Simulink, using the inverse dynamics and forward dynamics block, to control the motion of the moving platform. Actuator forces were determined by implementing vertebrate mastication trajectory as inverse dynamics using Matlab/Simulink and Visual Nastran. Results from the inverse dynamic simulations and motion control of the Stewart platform show that the Stewart platform enables the mastication motion to be reproduced.