-- Multi-Modal Input Environment Using Speech, Mouse and Key-Board --
Tetsunori KOBAYASHI
School of Science and Engineering, Waseda University
3-4-1 Okubo, Shinjuku-ku, Tokyo, Japan
koba@tk.elec.waseda.ac.jp
In this study, we regard ``Man-Machine Conversation'' as general interactive communication between human and machine including multimodel communication system. We investigate the framework to utilize speech input as powerful as possible in the general man-machine interface. Recently, the performance of the speech input system has been considerably improved and some demonstration systems have appeared. However they are not necessarily successful in appealing the effectiveness of speech input. This is mainly because the most of speech researchers concentrated just on using speech input and do not care for the total image of the desired interface system. To realize a useful speech conversation system, it is necessary to make it clear the desired role of speech input and utilize the characteristics. This year, we deal with a man-machine interface in a working environment using VDT. Task is a picture drawing tool. In this framework, the basic principles which are required for the useful input system are investigated. Input devices such as speech, mouse and key-board are verified from the view point of the proposed principles. Then, desired organization of the input system using these devices is considered, and prototype system which satisfies these condition is constructed. In conclusion, speech, mouse and key-board should basically fill their role to reduce the hand movement, but some alternative ways must be prepared to input each command for both of beginner and skilled. Moreover, it is pointed out that the importance of the natural guide which lead beginner to the skilled way.
Keywords: Spoken Dialogue, Multimodal Communication