Interactive and Continuous Collision Detection for Avatars in Virtual Environments

  • Authors:
  • Stephane Redon;Young J. Kim;Ming C. Lin;Dinesh Manocha;Jim Templeman

  • Affiliations:
  • University of North Carolina at Chapel Hill;Ewha Womans University, Korea;University of North Carolina at Chapel Hill;University of North Carolina at Chapel Hill;Naval Research Laboratory

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

Quantified Score

Hi-index 0.00

Visualization

Abstract

We present a fast algorithm for continuous collision detectionbetween a moving avatar and its surrounding virtualenvironment.We model the avatar as an articulated bodyusing line-skeletons with constant offsets and the virtual environmentas a collection of polygonized objects.Given theposition and orientation of the avatar at discrete time steps,we use an arbitrary in-between motion to interpolate thepath for each link between discrete instances.We boundthe swept-space of each link using a swept volume (SV) andcompute a bounding volume hierarchy to cull away linksthat are not in close proximity to the objects in the virtualenvironment.We generate the SV's of the remaining linksand use them to check for possible interferences and estimatethe time of collision between the surface of the SV andthe objects in the virtual environment.Furthermore, we usegraphics hardware to perform collision queries on the dynamicallygenerated swept surfaces.Our overall algorithmrequires no precomputation and is applicable to general articulatedbodies.We have implemented the algorithm ona 2.4 GHz Pentium IV PC with NVIDIA GeForce FX 5800graphics card and applied it to an avatar with 16 links, movingin a virtual environment composed of hundreds of thousandsof polygons.Our prototype system is able to detect allcontacts between the moving avatar and the environment in10-30 milliseconds.