Crowd simulation based on machine intelligent and biological behaviors

Abstract

Behavioral animation provides computer-generated creatures with instructions on how to react to external or internal stimulus, there by giving them autonomy. It has applications in entertainment and is also used as a tool in research. More recently, it has made its way into television and movie productions. However, its use in productions has its pros and cons. It can relieve animators from traditionally tedious tasks of animating creatures. But, with more autonomy, correct control of the creatures becomes unwieldy as high-level commands do not necessarily translate to the actual movements needed. The problem is compounded when crowds of autonomous creatures are to be generated. This project proposes and implements an animation system to generate and manage autonomous creatures. While applicable to different types of creatures, the emphasis is on a humanoid creature type. This creature is given an extensible action set with pre-defined actions, and an implemented behavioral module. This module is capable of computing mental attributes, handle miscellaneous and user-defined rules, and interpret terrain information. Finally, this module uses a condition-triggered method of defining behaviors, which addresses issues such as reusability, concurrency, and addition of states.