The Method for Supporting Children with Autism Spectrum Disorders using Paralinguistic Information

田中 宏季 (1051068)

Autism is a life-long developmental disability that prevents individuals from properly understanding what they see, hear, and otherwise sense. This results in severe difficulties in communication, and behavior. Children with autism spectrum disorders often do not understand the true intention of paralinguistic signals. We focused on laughter as one kind of paralinguistic information to support children with autism spectrum disorders. Laughter express interest and intention. This report presents the results of a preliminary analysis of the sounds of human laughs in a very large corpus of naturally-occurring conversational speech. Various types of laughter were categorized, and the analysis of their acoustic features forms the core of this report. Formant parameters ware compared for each call type within a laughter bout. The report also attempts to determine the contribution of individual acoustic features in the classification. Furthermore for children with autism spectrum disorders who can not speak, expressive speech synthesis including sample laughs was implemented. We developed a systematic speech synthesizer NOCOA (NOnverbal COmmunication for Autism), which aim to support nonverbal communication for Individuals with autism spectrum disorders. In this report, we explain the result and the benefit in experiment on NOCOA.