| Virtual reality technology not only offers users immersion but also is an effectiveway to reduce the adverse effects on the cognition of operators time delay brings inteleoperation system based on network. Manipulator and dexterous hand are actuators ofremote end in teleoperation. Dexterous hand as a multi-DOF tool is installed at the end ofmanipulator to execute intricate operations. Using data glove to drive hand is effective toavoid the shortcomings of computational complexity and poor interaction of the way ofanalysis. Collision detection is an important factor to ensure safe operation, make acorrect decision and improve immersion in the teleoperation system for dexterous handwith higher real-time human-computer interaction and security.The characteristics of3D modeling methods are firstly summarized. For theproject’s need, solid model of every object was built with3D modeling software. Withfurther development, posture information was extracted and grid models were created.The mechanism for creating and managing scene of Open Inventor was elaborated. Onthis basis, the virtual scene was constructed.The kinematics modeling of the whole scene was built after analyzing DOFconstraints and kinematics relationship. Kinematics analysis for manipulator was carriedout with the way of D-H matrix. It is more complicated to conduct kinematics analysisfor the dexterous hand with many joints and control it, so we gain the information of thefinger joints of the operator with CyberGlove and then drive dexterous hand to move.After the analysis of bounding volume hierarchies and parallel processingtechnology, a parallel collision detection algorithm was put forward which combines theadvantages of different bounding boxes and multicore programming.A virtual reality simulation platform for teleoperation was established with VC++6.0and Coin3D. The examples fully illustrate the methods above are correct and have auniversal practical value for simulation in teleoperation. |
PDF Full Text Request