Semantic Crowds: Reusable Population for Virtual Worlds

作     者：Nick Kraayenbrink Jassin Kessing Tim Tutenel Gerwin de Haan Fernando Marson Soraia R Musse Rafael Bidarra 

作者机构：Computer Graphics and Visualisation Group Delft University of Technology Delft The Netherlands Virtual Humans Simulation Lab Pontifícia Universidade Católica do Rio Grande do Sul Porto Alegre Brazil 

出 版 物：《Procedia Computer Science》 

年 卷 期：2012年第15卷

页      面：122-139页

主　　题：Crowd specification Reusable crowds Semantic virtual worlds 

摘      要：Recent advances in crowd simulation techniques have led to realistic agent and group behavior through elaborate behavioral models, complex motion planning algorithms and impressive physics systems. As many crowd simulation solutions typically target only specific types of environment and scenario, a variety of special-purpose methods and systems has emerged that are hard to re-configure and re-use in other contexts. Solving this situation demands a higher-level approach that takes re-use and configuration of crowds as a priority, for adequate application in a broad variety of scenarios, virtual environments and inter- action with the entities present in that environment. In this article we propose semantic crowds , a novel approach that allows one to re-use the same crowds for virtually any environment, and have them use the objects available in it in a meaningful manner, without any modification. To have the agents autonomously interact within any virtual world, we minimize in them the information relative to what objects do and how to use them. Instead, that information is stored in the objects themselves, which the agents can then query, based on what they plausibly want to achieve. To facilitate creating such crowds, we devel- oped an interactive crowd editor that provides high-level editing parameters for defining crowd templates. We illustrate the flexibility of semantic crowds by means of two cases, in which we let the same crowd populate quite differently configured airport terminal environments. These examples also highlight that this modular approach easily combines with your custom implementations of agent behavior model and/or motion planner.