In current CAD software the process of assembly modeling is hindered by a large number of separate rotation and translation actions necessary, especially in case of larger assemblies. Additionally matching faces, edges or points must be selected by clicking to define the appropriate constraint. In contrast to that, the process of assembling two normal sized physical parts in the real world seems to be rather simple. That is because we know how to grasp and move objects with our hands intuitively from our everyday experience. The idea behind this contribution is to enable the product developer to assemble CAD parts in a virtual environment through natural finger interaction like in reality. Therefore we present an overall method that combines the natural finger interaction with virtual objects and the insertion of constraints between rotationally symmetric CAD parts. The developed algorithms identify matching surfaces on the basis of the geometry as well as position and orientation of the parts in 3D space. This paper highlights the method to use a combination of real-time physics simulation and a heuristic approach to achieve an intuitive interaction interface. Additionally, we describe the detection algorithms developed to find assembly relationships between rotationally symmetric CAD parts without prior constraint definition. We also present a prototype system to demonstrate the functionality of the overall method. Furthermore, challenges for future research, such as extending the functionality of the detection algorithms on additional part types, like non-rotationally symmetric shapes, are discussed.

This content is only available via PDF.
You do not currently have access to this content.