Scalable Shared Virtual Worlds in the Browser

With the advancements in 3D graphics capabilities of commodity hardware and the evolution of the Web into a ubiquitous application platform, there is a renewed interest in creating Web applications featuring advanced 3D graphics. We focus on declarative 3D technology like XML3D that promises to bring the technology to the millions of Web developers and thus can truly scale with the Internet.

This project focuses on technology required to create scalable, real-time, visually rich, and complex virtual worlds (VWs) based on Web technologies.

We focus our research in particular on scalable client side technology. Here we build upon our declarative dataflow processing technology Xflow, which is currently used for character animations only. We will further extend Xflow to be applicable for surface shading, post/image processing as well as general optimization. As a result, we can use Xflow for visually rich virtual worlds with many dynamic objects or virtual characters as well as input processing, such as AR and gesture recognition. In collaboration with Intel Labs, we plan to tightly integrate River Trail into our implementation for parallelized JavaScript.

In addition, we include technology for scalable communicating with other VW clients and servers but are not addressing the server technology for VWs. Instead, here we plan to collaborate with other projects such as Stanford’s Sirikata and Intel Labs Virtual World Infrastructure research project.


Project team

Principal Investigator
Felix Klein

Kristian Sons
Johannes Ebersold
Sergiy Byelozyorov
Stephan Herhut