Title
Impact of Animal-Like Features on Emotion Expression of Robot Head EDDIE
Abstract
This paper investigates the influence of animal-like features (ears of a dragon lizard and comb of a cockatoo) and their pose variations on the quality of emotion expression of the robot head EDDIE. User studies are conducted presenting emotional expressions composed of the six basic emotions (based on Facial Action Coding Units (FACS) action units) in combination with the animal-like features under various inclination angles to human participants in randomized order. The results show that (i) the animal-like features have a significant influence on emotion expression, (ii) the expressed emotions are shifted in affective space dependent on the inclination angles of the animal-like features, (iii) these shifts are not uniform and differ in orientation for each of the expressed emotions, and (iv) for some emotions ground truth is matched better with the animal-like features in addition to FACS action units. The animal-like features, thus, help to improve the emotion expression of EDDIE, which is particularly beneficial because some action units have not been realized well or not realized at all. This study is a successful demonstration how expressive robot heads with rather technical design can benefit from additional expressive features, e. g., animal-like or fictionally designed, which are not based on FACS, in order to improve emotion expression. (C) Koninklijke Brill NV, Leiden and The Robotics Society of Japan, 2010
Year
DOI
Venue
2010
10.1163/016918610X501309
ADVANCED ROBOTICS
Keywords
Field
DocType
Emotions,robotics,emotion expression,facial expressions,human-robot interaction
Control theory,Emotion classification,Cognitive psychology,Coding (social sciences),Emotional expression,Artificial intelligence,Human–robot interaction,Computer vision,Facial expression,Engineering,Robot,Affect (psychology),User studies
Journal
Volume
Issue
ISSN
24
SP8-9
0169-1864
Citations 
PageRank 
References 
5
0.52
6
Authors
3
Name
Order
Citations
PageRank
Kolja Kühnlenz143040.87
Stefan Sosnowski216012.94
Martin Buss31799159.02