Page 184 - Designing Sociable Robots

breazeal-79017 book March 18, 2002 14:11

Facial Animation and Expression 165

Open
stance

accepting
Low
Negative arousal
valence sorrow
unhappy
anger tired
soothed
alert calm
content Positive
valence
fear joy
surprise
High disgust
arousal
stern
Closed
stance
Figure 10.3
The affect space consists of three dimensions. The extremes are: high arousal, low arousal, positive valence,
negative valence, open stance, and closed stance. The emotional processes can be mapped to this space.

Generating Emotive Expression
Kismet’s facial expressions are generated using an interpolation-based technique over a
three-dimensional space (see figure 10.3). The three dimensions correspond to arousal,
valence, and stance. Recall in chapter 8, the same three attributes are used to affectively
assess the myriad of environmental and internal factors that contribute to Kismet’s affective
state. I call the space defined by the [A, V, S] trio the affect space. The current affective
state occupies a single point in this space at a time. As the robot’s affective state changes,
this point moves about within this space. Note that this space not only maps to “emotional”
states (e.g., anger, fear, sadness, etc.) but also to the level of arousal as well (e.g.,
excitement and fatigue). A range of expressions generated with this technique is shown
in figure 10.4. The procedure runs in real-time, which is critical for social interaction.
The affect space can be roughly partitioned into regions that map to each emotion process
(see figure 10.3). The mapping is defined to be coarse at first, and the emotion system is
initially configured so that only limited regions of the overall space are frequented often. The
intention was to support the possibility of “emotional” and expressive development, where
the emotion processes continue to refine as secondary “emotions” are acquired through
experience and associated with particular regions in affect space with their corresponding
facial expressions.