Robust Joint Visual Attention for HRI Using a Laser Pointer for Perspective Alignment and Deictic Referring
In Human Robot Interaction (HRI), it is a basic prerequisite to guarantee joint attention, also known as shared attention, to get a proper coordination of the involved agents. A particular and important case of joint attention is joint visual attention, also referred to as perspective taking alignment, in which both the human agent and the robot must align their corresponding visual perspectives to look at the same scene or object of mutual interest. In this paper we present experimental work on the alignment of the visual perspectives of a humanoid-like robot and a human agent by means of a laser pointer used as a deictic or pointing device by both agents. We have developed experimental work to validate our proposed method using a scenario based on “I spy” game. After a brief discussion of joint visual attention, we introduce the humanoid-like robot specifically built for our experiments and afterwards we discuss the results obtained in the above-mentioned scenario. We would like to emphasize that for this scenario the human agents and the robot use limited linguistic words to facilitate coordination. These verbal exchanges are based on a common language (a lexicon plus grammar rules) for both humans and robots.
KeywordsHuman Robot Interaction Joint visual attention Deictic Humanoids
- 1.Bruner, J.: Nature and uses of immaturity. In: Conolly, K., Bruner, J. (eds.) The Growth of Competence, pp. 11–45. Academic Press, Cambridge (1974)Google Scholar
- 3.Butterworth, G.: The ontogeny and phylogeny of joint visual attention. In: Whiten, A. (ed.) Natural Theories of Mind, pp. 233–251. Blackwell, Oxford (1991)Google Scholar
- 5.Perzanowski, D., et al.: Integrating natural language and gesture in robotics domain. In: Proceedings of the IEEE International Symposium Intelligent Control, Piscataway, NJ, pp. 144–149 (1998)Google Scholar
- 7.Fuentes Brea, J.P., Maravall, D., de Lope, J.: Entropy-based search combined with a dual feedforward-feedback controller for landmark search and detection for the navigation of a UAV using visual topological maps. In: Armada, M.A., Sanfeliu, A., Ferre, M. (eds.) ROBOT 2013. AISC, vol. 253, pp. 65–76. Springer, Cham (2014). doi: 10.1007/978-3-319-03653-3_6 CrossRefGoogle Scholar
- 9.Maravall, D., Lope, J., Fuentes, J.P.: Visual bug algorithm for simultaneous robot homing and obstacle avoidance using visual topological maps in an unmanned ground vehicle. In: Ferrández Vicente, J.M., Álvarez-Sánchez, J.R., de la Paz López, F., Toledo-Moreo, F.J., Adeli, H. (eds.) IWINAC 2015. LNCS, vol. 9108, pp. 301–310. Springer, Cham (2015). doi: 10.1007/978-3-319-18833-1_32 CrossRefGoogle Scholar