Toward More Expressive Speech Communication in Human-Robot Interaction

It is well known that speech communication is a very important segment of human-robot interaction. The paper presents our experience from the project “Design of Robots as Assistive Technology for the Treatment of Children with Developmental Disorders”, with focus on the development of more expressive dialogue systems based on automatic speech recognition (ASR) and text-to-speech synthesis (TTS) in South Slavic languages. The paper presents the most recent results of our research related to the development of expressive conversational human-robot interaction, specifically in the field of conversion of voice and style of synthesized speech based on a new generation of deep neural network (DNN) based speech synthesis algorithms, as well as the field of emotional speech recognition. The development of dialogue strategies is described in more details in the second part of the paper, as well as the experience in their clinical applications for treatment of children with cerebral palsy.