Optimizing avatar environmental update in shared virtual reality environments

A major problem with the current generation of shared 3D virtual reality game-like environments on the Internet is that update of the virtual environment during avatar movement often lags considerably behind the scene which the avatar's user perceives. The result is an unsatisfying, artificial experience, in which the user must halt avatar movement until the new scene slowly assembles before the avatar and sounds pick up, or allow the avatar to continue moving and risk bumping into objects that are revealed later when the view fully assembles. This paper describes an algorithm for optimizing environmental update based on a caching strategy for prefetching environmental data around a stationary avatar and a statistical model predicting avatar movement that is used for determining when and where to start updating the environmental data when the avatar starts moving.