What could assistance robots learn from assistance dogs?

Gácsi, Márta and Szakadát, Sára and Miklósi, Ádám (2013) What could assistance robots learn from assistance dogs? In: Bionetics 2012: 7th International Conference on Bio-Inspired Models of Network, Information, and Computing Systems., 2012.12.10-2012.12.11., Lugano, Svájc.

Preview

Text Gacsi et al Lugano.pdf Download (261kB) | Preview

Abstract

These studies are part of our broader project that aims at revealing relevant aspects of human-dog interactions, which could help to develop and test robot social behaviour. We suggest that the cooperation between assistance dogs and their disabled owners could serve as a model to design successful assistance robot–human interactions. In Study 1, we analysed the behaviour of 32 assistance dog–owner dyads performing a fetch and carry task. In addition to important typical behaviours (attracting attention, eye-contact, comprehending pointing gestures), we found differences depending on how experienced the dyad was and whether the owner used a wheel chair or not. In Study 2 we investigated the reactions of a subsample of dogs to unforeseen difficulties during a retrieving task. We revealed different types of communicative and displacement behaviours, and importantly, dogs showed a strong commitment to execute the insoluble task or at least their behaviours lent a “busy” appearance to them, which can attenuate the owners’ disappointment. We suggest that assistant robots should communicate their inability to solve a problem using simple behaviours (non-verbal vocalisation, orientation alternation), and/or could show displacement behaviours rather than simply not performing the task. In sum, we propose that assistant dogs’ communicative behaviours and problem solving strategies could inspire the development of the relevant functions and social behaviours of assistance robots.

Actions (login required)

Edit Item