Multimodal human-robot interaction with Chatterbot system: Extending AIML towards supporting embodied interactions

Abstract

The research objective of this work is to realize multimodal human-robot interaction based on light-weight Chatterbot system. The dialogue system is integrated into SIGVerse system with immersive multimodal interfaces to achieve interaction in an embodied virtual environment. To validate the feasibility of the proposed design, the actual AIML implementations are described to illustrate (a) Gesture Inputs, (b) Emotional Expressions, (c) Robot Interactive Learning, and (d) Interactive Learning towards Symbol Grounding in multimodal human-robot interaction.

Integration of work sequence and embodied interaction for collaborative work based human-robot interaction

“…3) Verbal Communication: The task nodes' descriptions are written in human language and can be used by robot as verbal support (text-to-speech) to human, together with the dialogue engine developed in [3].…”

Embodied and multimodal human-robot interaction between virtual and real worlds

“…3 shows a "Clean Up" task, where the robot is conversing verbally with the human to obtain the information of the objects and the surroundings. Conversational intelligence and dialogue management system [5] are developed to understand the context (including the temporal and spatial information of the objects and environment) in order to deduce the full meaning from the partial verbal instruction giving by the human. In the Okonomiyaki collaborative cooking task (Fig.…”