A framework to investigate behavioural models

Abstract

This paper presents a framework to investigate behavioural models of multiple agents. The framework is designed to deal with problems commonly found in behavioural animation systems, most of which are caused by excessive coupling between the system modules. Hence, the framework is designed to be modular, exible and extensible. Behavioural animation models based on this framework are clearly divided in modules that can be independently designed and developed. Furthermore, these modules can be easily substituted, modi ed and reused. We present modules related to crowd behaviour, intelligent camera, and virtual environment and how these are integrated using the Python programming language. We also discuss visualization aspects, which are addressed by yet another module.