Publication: A framework for crowd simulation based on the JMonkey game engine
Loading...
Official URL
Full text at PDC
Publication Date
2017
Authors
Advisors (or tutors)
Editors
Journal Title
Journal ISSN
Volume Title
Publisher
Citations
Exportar
Abstract
La simulación de multitudes juega un papel crucial cuando se trata del desarrollo de entornos inteligentes. La mayoría de los investigadores desarrollan simulaciones usando motores de juegos comerciales a través de los editores que éstos proporcionan. Esto di culta el poder realizar una experimentación profunda sobre simulaciones de multitudes, y fuerza que la línea de investigación deba atenerse al paradigma de desarrollo propuesto por la herramienta. El objetivo principal del trabajo desarrollado es la contribución de un simulador de multitudes basado en 3D, con una arquitectura modular y extensible, adecuada para la experimentación con simulaciones de multitudes. Este framework se centrará de forma especial en la navegación y la coordinación de multitudes sobre un modelo realista del entorno, capaz de reproducir situaciones del mundo real. El simulador incluye implementaciones de algoritmos conocidos para el movimiento de multitudes, integrando también implementaciones de terceros.
El trabajo tiene en cuenta la necesidad de representaciones visualmente convincentes de la simulación más allá de las representaciones 2D, utilizadas regularmente en la literatura. Para ello, se contribuye con extensiones a herramientas de terceros que permiten importar texturas, animaciones y mallas que mejoran la calidad de la simulación.
El desempeño de la simulación se demuestra en un caso de estudio donde el desafío es encontrar una población cuyo comportamiento, dentro del simulador, reproduce un determinado tráfico entrante / saliente medido en áreas específicas de un edificio.
Este trabajo ha sido financiado por el proyecto MOSI-AGIL (S2013 / ICE-3019) a través de la Gobierno de la Comunidad de Madrid y Fondos Estructurales Europeos (FEDER).
Crowd simulation plays a crucial role when it comes to the development of Smart Environments. Most researchers develop simulations using commercial game engines through the editors they provide. This prevents from a deep experimentation with the problems of crowd simulation and enforces to stick to the development paradigm of the tool. The main purpose of the work developed is to contribute with a 3D based crowd simulator framework that runs with a modular, extensible architecture suitable for experimentation with crowd simulation problems, such as steering or collision avoidance. This simulation framework will have an special focus on the navigation and crowd coordination within a model of a physical environment having the capacity of reproducing real world situations. The simulation framework includes custom implementations of known algorithms for steering and path smoothing and reuses implementations for collision avoidance. The work takes into account the need of visually convincing representations of the simulation beyond the 2D regularly used in the literature. For this aim, it contributions with extensions to third party tools that allow to import textures, animations and meshes that improve the quality of the simulation. The performance of the simulation is demonstrated in a case study where the challenge is to find a population whose behavior, within the simulator, reproduces a particular incoming/outgoing traffic measured in specific areas of a building. This work has been funded by the MOSI-AGIL (S2013/ICE-3019) project through the Government of the Region of Madrid, and European Structural Funds (FEDER).
Crowd simulation plays a crucial role when it comes to the development of Smart Environments. Most researchers develop simulations using commercial game engines through the editors they provide. This prevents from a deep experimentation with the problems of crowd simulation and enforces to stick to the development paradigm of the tool. The main purpose of the work developed is to contribute with a 3D based crowd simulator framework that runs with a modular, extensible architecture suitable for experimentation with crowd simulation problems, such as steering or collision avoidance. This simulation framework will have an special focus on the navigation and crowd coordination within a model of a physical environment having the capacity of reproducing real world situations. The simulation framework includes custom implementations of known algorithms for steering and path smoothing and reuses implementations for collision avoidance. The work takes into account the need of visually convincing representations of the simulation beyond the 2D regularly used in the literature. For this aim, it contributions with extensions to third party tools that allow to import textures, animations and meshes that improve the quality of the simulation. The performance of the simulation is demonstrated in a case study where the challenge is to find a population whose behavior, within the simulator, reproduces a particular incoming/outgoing traffic measured in specific areas of a building. This work has been funded by the MOSI-AGIL (S2013/ICE-3019) project through the Government of the Region of Madrid, and European Structural Funds (FEDER).
Description
Máster en Ingeniería Informática, Facultad de Informática, Departamento de Ingeniería del Software e Inteligencia Artificial curso 2016-2017