Research And Application Of The Dexterous Hand Grasp With Force Feedback In Virtual Environments

Intuitive and natural interaction between users and virtual objects is a hot research issue in the field of virtual reality, since it is very important to enhance users' immersion experience as well as their percept in virtual environments (VE). But it is also a big challenge, because VE is unable to provide physical constraints necessary for interaction as the real world does. In this thesis, we propose an approach to grasping virtual objects via a dexterous hand as an avatar in VE to help users interact with VE intuitively and naturally.At first, hand's grasp trajectory is utilized to detect collisions between a dexterous hand and virtual objects and then the corresponding grasp posture is estimated through measuring fingers' automatic contact conditions.Subsequently, two conditions are given to identify whether the dexterous hand could grasp and manipulate the object. One is based on collision results between the dexterous hand and objects, and the other is by assessing dexterous hand's grasp ability to resist external force according to its contact positions on the object.At the same time, users are also provided with haptic feedback during their operations in VE to enhance their immersion experience. Aiming to different states of the dexterous hand, different force computation models are adopted.At last, a virtual assembly prototype system is developed. The core of the system is to turn assemblage operations into deassemblage ones, since we believe if a product could be deassembled, it could also be assembled. Based on this principle, the sequence and route of assemble could be easily deducedby recording the process of deassemblage.