You are here:

Multi-user virtual learning environments: Design, implementation, and avatar behavior

, George Mason University, United States

George Mason University . Awarded


Multi-user virtual environments (MUVEs) have been studied, developed, and used in military and industrial team training, collaborative design and engineering, and multiplayer games over the years. However, little effort has been put into developing MUVEs for teaching and learning purposes. This dissertation concentrates on the design and implementation of multi-user virtual learning environments.

First, as a team member, I designed and implemented a multi-user virtual learning environment system, MUVEES (Museum-Related Multimedia and Virtual Environments for Teaching and Learning Science), a National Science Foundation funded research project. The novel architecture, hierarchical multiple-server architecture, allows MUVEES to cover a large geographical area, handled a large number of concurrent participants, functioned as a virtual world container, and dealt with special teaching and learning requirements. This architecture has workload balancing capability, high reliability, and good scalability. The MUVEES second version is now under test in the School of Education at both Harvard University and George Mason University. The first version has been running since Fall 2001 in many middle schools.

Second, I proposed a new mechanism to allow participants to navigate complex virtual worlds instantly when they enter them. When participants enter other virtual worlds, their computers are usually frozen until this loading procedure is done. This phenomenon makes participants wait for a tedious period (usually a couple of minutes) before they can view the new virtual world. I developed a novel mechanism to reduce this discomfort. Compared to the conventional loading procedure, experiments show that the extra time required by my mechanism is trivial. The primary benefit is that MUVEES participants are able to interact with their virtual worlds immediately when they enter them.

Third, I introduced a new approach to achieve human-like avatar behavior in virtual environments. Having human-like avatars in virtual environments significantly increases the realism of virtual environments. This is more important in multi-user virtual learning environments, as the avatars are used for participants to communicate and collaborate with each other. Traditional methods require intensive calculations and a significant amount of rendering resources. The image-based human-like avatar behavior approach I proposed uses pre-generated avatar images to implement avatar behavior in virtual environments. This is the first effort to apply image-based rendering techniques onto active avatars. This approach also allows us to put real actors into virtual worlds.

Finally, I proposed a novel method, the hybrid approach, to speed up rendering active avatars. My approach is different from any other existing methods. The heart of this approach is that the cached segment images from the 3D avatar geometric model are reused to generate subsequent animation frames within bearable visual error. The animated behavior is changeable through modifying parameters or experimental data. Compared to existing rendering methods, my implementation and statistics show that the proposed method significantly reduces the rendering time. Taking into account of the appearance of multiple avatars in a shared virtual world, the overall rendering performance will be significantly improved.

In summary, this dissertation addresses the design and implementation of multi-user virtual learning environments and explores new methods to achieve human-like avatar behavior in virtual environments. The new approaches and algorithms proposed in this dissertation are more efficient than the similar existing results, and can be used to improve the performance of multi-user virtual environments.


Yang, Y. Multi-user virtual learning environments: Design, implementation, and avatar behavior. Ph.D. thesis, George Mason University. Retrieved October 20, 2019 from .

This record was imported from ProQuest on October 23, 2013. [Original Record]

Citation reproduced with permission of ProQuest LLC.

For copies of dissertations and theses: (800) 521-0600/(734) 761-4700 or