Multimodal dialogue system with NAO and VoiceXML dialogue manager

Abstract

The proposed paper describes a multimodal interactive system based on NAO humanoid robot with an external dialogue management module VoiceON. System can be controlled by voice. NAO produces speech and gestures as a response to the user inputs. Presented multimodal system uses built-in speech recognition and speech synthesis modules adapted to Slovak language, which was originally not supported. Moreover, system accepts VoiceXML dialogue applications for all supported languages. To manage dialogue interaction, a previously developed VoiceXML dialogue manager was adopted. A new module for generation of multimodal output (speech and gestures) was designed, which enables NAO to produce gestures together with speech in several modes.