Open Source System Integration Towards Natural Interaction with Robots

Abstract

Speech is an intuitive way to interact with social robots: spoken language dialogues can help users to express their intents in a natural and flexible manner. In recent years, there has been remarkable progress in artificial intelligence related to spoken dialogue technology, including speech recognition and natural language processing. In this paper, we present the integration of the open source speech recognition, natural language processing, and dialogue management components into a robot software platform, and also report on a preliminary experiment of the integrated system using real users. Gesturing of the robot, which is also important in human-robot interaction, is combined with the spoken content of the robot utterance and included in the dialogue management component. As the dialogue domain we chose mealtime discussions on food and recipes, since spoken communication with a companion robot in such scenarios is considered natural and useful.