Dynamic Modeling And Mechanism Performance Analysis Of A Parallel Haptic Device

Haptic devices are the key components in the teleoperation and virtual realitysystems. As part of the 863 projects of"Study of a Robot Assisted Orthopedic Tele-operation System"and"Study of the Key Technologies of Telemedical Robots and Sy-stems", This thesis builds the kinematic and dynamic models of a 6-DOF 3-subchainhaptic device, analyzes its workspace and dexterity, and optimizes its kinematics pa-rameters.Considering that the haptic device is a serial/parallel hybrid mechanism, the ki-nematic analysis is firstly performed after the mechanism analysis. The inverse ki-nematic models are built using the coordinate transformation method. The Newton-Raphson method is used to solve the nonlinear equations in the forward kinematics.And the velocity and acceleration mapping equations are derived through the vectorchain method.Based on the kinematic model, the mathematic model of the inverse dynamicsis built by the Kane equation, since conventional dynamic modeling methods willresult in complex derivation and computation. And the forward dynamic model isbuilt using Adams. For two motion instances of the top plate, the mathematic model isused for obtaining the six input torques, which are then imported to the Adams modelto actuate the corresponding links. By comparing the motion of the Adams model withthe given motion, the mathematic model is validated. Then a simple dynamic modelsolely considering the mass of the top plate is built for comparison.In the workspace analysis, a new set of Euler angles called T&T angles is utilizedto represent the orientation of the top plate. The spherical scanning search method isused to obtain complete position workspace. The position and orientation workspaceof the haptic device is analyzed based on those methods. Besides, a novel position-orientation-coupled workspace called position workspace with a specified orientationcapability requirement is presented as a more appropriate representation of workspacecapability.In order to effectively analyze the dexterity, addressing the unit–inhomogeneityissue in the Jacobian matrix, a new general normalization method is proposed. Andthis method is used to analyze the local dexterity of the haptic device. Then the genetic algorithm tool in Matlab is used for the optimization of the kinematic parameters, withworkspace and compactness as the optimization constraints, and global dexterity asthe optimization objective. The result is validated by comparing the dexterity of themechanism before and after the optimization.