This contribution introduces a new framework for assembly modeling in Virtual Reality. Aiming at an easy instructability of the virtual environment, the framework provides a task-level interface which directly maps logical assembly commands to corresponding changes in the geometry scene. For example, the visual assembly of two parts is achieved given only a single command 'connect (a,b)'. This is in contrast to the assembly modeling style of conventional CAD systems which forces the designer to break down each conceptual assembly task into a series of lower-level subtasks. The proposed framework consists of two parts: (1) A knowledge-based model of connection-sensitive part features ("ports") and the connections between them; and, (2) a set of algorithms that define the task-level interface for assembly, disassembly, and adjustment the real-time requirements of virtual environments. At the user interface, both direct manipulation and directive interfaces, e.g. based on natural language instructions are supported. A family of implemented VR-systems, including AVE and Internet-based applications, demonstrates the feasibility of the approach.
展开▼