Yiqun Chen and Ian D. Bishop
Multi-Agent System, Virtual Environment, 3D Modelling , Visualization, Visitor Simulation, Game Engine
In the context of urban park management, knowledge of visitor recreational activities is imperative to planners to design and manage their lands and facilities effectively. A well calibrated Multi-Agent System (MAS) for visitor simulation can offer planners a better understanding either in the current situation or in any proposed scenarios. In addition, by reconstructing scenes based on Geographic Information System (GIS) data and detailed modeling, 3D Virtual Environments (VE) provide an intuitive and vivid way to mimic the real world and fundamentally extend agent perception and exploring abilities. This paper reports the approaches of using a multi-agent model in a virtual environment to simulate visitor activities in Australian Garden (AG), Royal Botanic Gardens Cranbourne (RBGC). To understand how visitors interact with landscapes, an iPhone-based visitor survey is conducted to collect on-site data; then a 3D virtual environment is constructed using GIS data based on Torque Game Engine Advanced (TGEA); a navigation mesh and attraction point structure are introduced to broaden agent movements; finally, equipped with vision depth detection, attraction point reasoning and landscape aesthetic ability, an intelligent agent model is proposed.
Important Links:
Go Back
