分化感知建模研究  

Study on Modeling of Differential Perception

作　　者：郭小明[1] 崔建明[1] 何恭贺[1] 王治国[1] 

机构地区：[1]桂林理工大学信息科学与工程学院,广西桂林541004

出　　处：《计算机系统应用》2010年第6期37-39,共3页Computer Systems & Applications

摘　　要：针对计算机智能角色制作中感知信息难以准确提取的问题提出了"感知分化"的概念并根据该思想设计了基于蛇类的感知模型。而且在感知信息的提取和处理之间引入了信息预处理过程,实现对"分化"的感知信息"汇总"过程。此过程的信息处理依托于人工神经网络的数学模型,能够很好的根据"分化"的感知信息重构角色关注的对象。基于上述思想设计的智能角色具有更快的实时响应,基本能够准确的完成任务。To address the problem of extracting perception information difficultly in making intelligent digital role, we propose the concept of differential according perception and according to it design a perception model based on the snake. A new process ofpretreatment of information has been inserted between obtaining information and information processing, which will inosculate all input information. The pretreatment relies on the mathematic model of artificial neural net, which could better reconstruct the object focused by digital role with the differential information. The intelligent role designed with our policy has faster real-time response, and can finish the task accurately.

关 键 词：感知分化 信息预处理 人工神经网络 重构 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]