建立Avatar感知能力与自主行为选择模型的新方法  

New perception and autonomous behavior modeling approach for Avatar

作　　者：陈中育[1] 周晴红[1] 朱响斌[1] 

机构地区：[1]浙江师范大学信息科学与工程学院,浙江金华321004

出　　处：《计算机应用》2006年第4期776-779,共4页journal of Computer Applications

基　　金：浙江省自然科学基金重点重大资助项目(Z603262)

摘　　要：在人脸重构系统与智能虚拟环境系统的研究基础上,提出了一种基于Agent的建立虚拟人感知能力与行为选择模型的新方法。该方法把虚拟人(Avatar)当成一个Agent,它在虚拟场景中能够识别三维物体以及利用传感器计算物体的运动方向。Avatar的行走采用骨架模型来进行仿真,并且其行走行为由七个基本动作行为组成。实验结果显示新方法是具有较好的效果。Based on the research of 3D facial model reconstruction and intelligent virtual entrainment, a new perception and autonomous behavior modeling approach for Avatar was proposed, in the new method, the Avatar was considered as an Agent, which could recognize 3D objects in the environment and calculate the direction of the moving objects by using the Avatar＇s perception sensor. Avatar walk behavior was simulated by skeleton model and denoted by seven motion sequences, Experiment results show that the new approach is more efficiency.

关 键 词：智能虚拟环境 行为动画 智能感知 行为决定 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术] TP391.9[自动化与计算机技术—计算机科学与技术]