Abstract: Multimodal Interfaces employ many kinds of interactive devices, such as speech input/output, pointing, gazetracking, and tactile and force feedback devices, make them cooperate with each other and take the advantage of their complementarily so as to facilitate the communication between user and computer, increase the efficiency of interaction, enhance its naturalness and ultimately make it possible for computer-naive people to interact with computers using their everyday skills. This paper summarizes part of the work completed in the key project Multimodal User Interface Research supported by National Natural Science Foundations of China. First, we described its concepts, advantages and some terminologies, then proposed an Architecture and integration for Task Oriented Multimodality (ATOM) model, a Hierarchical Multimodal Interface Specification (HMIS) method, and a new interface paradigm (VIR-Virtual Information Room). Finally, two case studies developed under way and conclusion are given.

Key words: user interface, multimodal interface, user interface model, user interface specification method, user interface paradigm, multimodal intergration, user interface, multimodal interface, user interface model, user interface specification method, user interface paradigm, multimodal intergration

摘要： 多通道界面使用多种交互设备(语音输入输出装置、直接指点装置、视线跟踪装置、触觉和力量反馈装置)，通过多种交互方式的协作，利用它们之间特性上的互补，方便用户传达交互意图和理解计算机的输出，提高交互效率，增进交互的自然性，目的是要让普通用户能够以日常技能来使用计算机。本文以“多通道用户界面研究”为背景，总结了北京大学完成的部分工作。首先对多通道界面的概念、优势和一些术语进行了说明，然后提出了面向任务的多通道界面结构和整合模型(ATOM)，层次化的多通道界面描述方法(HMIS)和新界面范式(VIR)。最后给出了两个多通道实例及结论。

关键词: 用户界面, 多通道界面, 界面模型, 界面描述方法, 界面范式, 多通道整合, 用户界面, 多通道界面, 界面模型, 界面描述方法, 界面范式, 多通道整合