多通道整合的相关问题及算法  

Some Questions and Algorithm of Multimodal Integration

作　　者：张宏超[1] 俸文[1] 周方[2] 孙亚民[1] 

机构地区：[1]南京理工大学计算机科学与技术系,南京210094 [2]中国电子科技集团第二十八研究所,南京210007

出　　处：《计算机工程》2004年第13期67-68,171,共3页Computer Engineering

基　　金：国家十五重点军事预研项目资助

摘　　要：多通道界面旨在充分利用一个以上的感觉和动作通道来捕捉用户的意向,提高人机交互的自然性和高效性,其核心问题就是通道整合的问题。该文针对该问题提出了一个基于分层的任务模型的整合算法,该算法处理中涉及到程序运行状态表示、任务结构的设计和交互原语相关性表示等问题。针对以上3个问题展开讨论,在此基础上得到最后的算法。The aim of multimodal interface is to get the users intent by more than one sense and action modals and to make the interaction of human an computer more natural and more effective. The key of multimodal interface is multimodal integration. In this article an integration algorithm based on task oriented and hierarchical multimodal integration modal is carried out. In this algorithm there are three questions: representation of programme state, desig of task structure and representation of the relativity of interaction primitive to be discussed. At last the algorithm is given.

关 键 词：多通道用户界面 有穷自动机 对象属性结构 交互原语 

分 类 号：TP39[自动化与计算机技术—计算机应用技术]