Unimodal and multimodal variants of voice output communication aids

点击翻译按钮获取中文摘要

Individuals with little or no functional speech, gesture, or writingdue to underlying physical disorder may instead use a computationaldevice, called a Voice Output Communication Aid (VOCA), to producesynthetic speech (synthesized or digitized).  This device affordsanother, albeit aided, mode of communication.  Empirical evidenceshows that this mode is a complement to, and not a replacement for,the other, unaided modes, such as vocalization, facial expression andgaze, and gestures of the hand, head and torso.  Thus, the evaluationand design of VOCAs must take into consideration not only the issue ofrate of output (e.g., words-per-minute) but also issues that relate tothe utility of the individual's entire repertoire of mode strategies,of which synthesized speech is but one component.The goal of this program of research is to develop better VOCAsthrough the development of software techniques and modelling toolsthat take into account this "global" context.  The research will focuson the contrastive evaluation of unimodal and multimodal interfacevariants, and the use of design principles based on informationtheory.  The formal and theoretical analyses will be validated byempirical studies.