The new wave of computer-driven entertainment technology throws audiences and game players into massive virtual worlds where entire cities are rendered in real time. Computer animated characters run through inner-city streets teeming with pedestrians, all fully rendered with 3D graphics, animations, particle effects and linked to 3D sound effects to produce more realistic and immersive computer-hosted entertainment experiences than ever before. Computing all of this detail at once is enormously computationally expensive, and game designers as a rule, have sacrificed the behavioural realism in favour of better graphics. In this paper we propose a new Collision Avoidance Level of Detail (CA-LOD) algorithm that allows games to support huge crowds in real time with the appearance of more intelligent behaviour. We propose two collision avoidance models used for two different CA-LODs: a fuzzy steering focusing on the performances, and a geometric steering to obtain the best realism. Mixing these approaches allows to obtain thousands of autonomous characters in real time, resulting in a scalable but still controllable crowd. © Springer-Verlag 2009.
CITATION STYLE
Paris, S., Gerdelan, A., & O’Sullivan, C. (2009). CA-LOD: Collision avoidance level of detail for scalable, controllable crowds. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 5884 LNCS, pp. 13–28). https://doi.org/10.1007/978-3-642-10347-6_2
Mendeley helps you to discover research relevant for your work.