Robots are increasinlgy being marketed for consumer applications (e.g. cleaning or entertainment). A natural language interaction is more and more expected to make them more appealing and accessible to the end user. Latest advances in speech technologies enable different and richer levels of interaction. In Human Robot Interaction, command understanding is the baseline for a natural interaction. More complex interactions require different levels of natural language processing capabilities, such as dialogue management.
In our group we deal with the Natural Language Understanding for Robotic platforms. In this area we investigate how state-of-the-art textual inference technologies can be employed and adapted for HRI systems. The final aim of this research is to propose a unifying framework able to capture semantic aspects as these are needed in the HRI area. We foster the idea that many problems tackled and solved in Natural Language Processing, e.g. Semantic Role Labeling (SRL) (Palmer et al., 2010), can be taken into account for HRI.
HuRIC: a Human Robot Interaction Corpus.
D.L.Chen,R.J. Mooney,Learning to interpret natural language navigation instructions from observations,in Proceedings of the 25th AAAI Conference on AI, 2011, pp. 859–865.
G.J.M. Kruijff, H.Zender, P.Jensfelt, H.I.Christensen, Situated dialogue and spatial organisation: What, where and why?, International Journal of Advanced Robotic Systems 4 (2).
J.Bos, T.Oka, A spoken language interface with a mobile robot, Artificial Life and Robotics 11(1) (2007)42