Analysis of parallel manipulator architectures for use in masticatory studies
Sathia Narayanan, Madusudanan
MetadataShow full item record
The goal of this work is to (i) quantitatively analyze the jaw motions using a variety of quantitative engineering tools and (ii) create/design a virtual jaw motion simulator based on parallel architecture manipulator that can reproduce these recorded jaw motions of vertebrates/animals accurately. Such an implementation could provide a test bed to quantitatively characterize the jaw motion based on "chewability index" factor for wide range of applications. To this end we will examine two sets of case studies--human jaw motion and canine jaw motion. In the first phase, we begin with initially developing the biomechanical models of a human and canine jaw in AnyBody. The human model is developed using elements from the AnyBody repository while the canine jaw was developed from scratch. For this purpose, the muscle model parameters as well as kinematic modeling of temporomandibular joint are ascertained by a real dog cadaver dissection. Detailed case studies were conducted to validate the inverse dynamics analyses and work envelope of both the models--human and labrador dog. We also artificially simulated the bite force in these biomechanical models as a part of this case study for further validation. In the second phase, we examine the different parallel architecture manipulator systems that would suit our desired application resulting in final selection of two manipulator configurations for our application-6 DOF R -U-S (revolute-universal-spherical with active revolute joints) and P -U-S (prismatic-universal-spherical with active prismatic joints) manipulators. Accurate kinematic models of these manipulators were developed in MATLAB and these were used to evaluate the manipulators based on Jacobian based measures. Kinematics of Parallel manipulators is, in general, modeled using simple loop closure technique and Jacobian matrices were derived using screw-theoretic modeling. We then completed a comprehensive parametric studies based on maximum workspace and force limits for P -U-S manipulator case. Motion capture trajectories for the jaw motion are obtained using the high speed SimiMotion motion capture system with the real subjects (humans and dogs). For ascertaining the model of simulator, we finally use the motion capture trajectories to drive the AnyBody as well as the parallel manipulator model and verify the workspace envelope to match with the jaw trajectories. Finally, we briefly discuss the dynamic modeling of the system for real-time with hardware-in-loop simulation and physical prototype implementation using rapid prototypes as well as cast dentitions.