|title:||Needs-based Companion Dog|
|author:||Jose Mar¡a Blanco Calvo|
|published in:||October 2007|
Master of Science thesis
Man-machine interaction group
Delft University of Technology
|PDF (5905 KB)|
This thesis describes a project carried out at the Man Machine Interaction research
group which is part of Mediamatics at the Delft University of Technology in the
Netherlands. Needs-based companion dog wants to open new ways of researching in the
interaction between robots and humans. In particular the project has focussed on creating a
complex personality model for a robotic companion dog.
The societal relevance of intelligent robots is increasing nowadays covering a wide range of applications: from entertainment robots such as conversational partners, soccer players, or companion dogs, to robots that provide independent living support for elderly users in basic activities such as mobility or household maintenance. Security tasks can also be performed by surveillance robots like watchdogs. Robots have shown their utility even in medical domains by aiding the diagnosis and therapy of diseases as autism.
After studying some existing models of artificial personalities, we propose an extension of the nPME model, constituted by needs, Personality, Mood and Emotions. Our main contribution is the use of needs (physiological needs, safety needs, and love & belonging needs) as the main forces that control coherent behaviour. We set the satisfaction of the needs always in combination with incoming external events as the basis to evaluate the coherence and realism of the behaviour. Personality and emotions also play an important role in the model so special effort has been put on them to improve the realism of the scene.
Technically we have used a robotic dog called AIBO (made by Sony) as the hardware platform, which was managed by a laptop through a wireless connection. The software for the main control of the system has been implemented using Java. The URBI platform was used to provide the physical control of the robot. The current implementation of the nPME model comprises two highly coupled expert systems built using JESS.
Several experiments were carried out to test the validity of our approach. Among them, we did a comparison with the commercial model of Sony, AIBO's mind software, and an user study with 24 people to test the realism and coherence of the model for a robotic companion dog. Although results suggest that further work should be done in improving reaction times and some particular actions, they also show that our model was successful in showing a convincing behaviour and coherent emotional responses as a robotic companion dog.