Improving Collision Detection in Distributed Virtual Environments by Adaptive Collision Prediction Tracking

• Authors: Jan Ohlenburg
• Venue: VR '04 Proceedings of the IEEE Virtual Reality 2004
• Year: 2004

Collision detection for dynamic objects in distributedvirtual environments is still an open research topic. Theproblems of network latency and available network bandwidthprevent exact common solutions. The Consistency-Throughput Tradeoff states that a distributed virtual environmentcannot be consistent and highly dynamic at thesame time. Remote object visualization is used to extrapolateand predict the movement of remote objects reducingthe bandwidth required for good approximations of the remoteobjects. Few update messages aggravate the effect ofnetwork latency for collision detection.In this paper new approaches extending remote object visualizationtechniques will be demonstrated to improve theresults of collision detection in distributed virtual environments.We will show how this can significantly reduce theapproximation errors caused by remote object visualizationtechniques. This is done by predicting collisions betweenremote objects and adaptively changing the parameters ofthese techniques.