Article,
Speech Centric Multimodal Interfaces for Mobile Communications Systems
Telektronikk, (2003)
Abstract
Natural conversational man-machine interfaces should support multimodal interaction. Generally, a multimodal system combines natural input modes such as speech, touch, manual gestures, gaze, head and body movements, and searches for the meaning of these combined inputs. The response can be presented by a multimedia system. The main focus of this paper is the technical aspects of implementing multimodal interfaces for mobile terminals. With the limited size and processing power of these terminals, we have restricted the functionality to speech centric multimodal interfaces with two input modes: speech (audio) and touch, and two output modes: audio and vision. That is, the input combines automatic speech recognition and a pen to click areas on the touch-screen, or pushing buttons on the small terminal. The output is either speech (synthetic or pre-recorded) or text and graphics.
