Fidelity Optimization in Distributed Virtual Environments

摘要

In virtual environment systems, the ultimate goal is delivery of the highest-fidelity user experience possible. This dissertation shows that is possible to increase the scalability of distributed virtual environments (DVEs), in a tractable fashion, through a novel application of optimization techniques. Fidelity is maximized by utilizing the given display and network capacity in an optimal fashion, individually tuned for multiple users, in a manner most appropriate to a specific DVE application. This optimization is accomplished using the QUICK framework for managing the display and request of representations for virtual objects. Ratings of representation Quality, object Importance, and representation Cost are included in model descriptions as special annotations. The QUICK optimization computes the fidelity contribution of a representation by combining these annotations with specifications of user task and platform capability. This dissertation contributes the QUICK optimization algorithms; a software framework for experimentation; and associated general purpose formats for codifying Quality, Importance, Cost, task, and platform capability. Experimentation with the QUICK framework has shown overwhelming advantages in comparison with standard resource management techniques.