Robotics Reference
In-Depth Information
to distinguish between the inflections associated with the full range of
Ortony/Clore/Collins emotions and even more.
An example of how people model one another's points of view, in
order to explain and predict each other's responses to various situations,
is the Affective Reasoner program developed by Elliott and Ortony, that
reasoned about the emotions of robots in a simulated world. The robots
simulated by these programs were given a simple emotional life consist-
ing of the 22 Ortony/Clore/Collins categories of emotion plus love and
hate, and they were endowed with approximately 1,200 different expres-
sions of emotion. Each simulated robot was given its own personality,
so that different robots could have different interpretations within their
own worlds, and exhibit different response tendencies.
In order for one robot to understand how another robot or a hu-
man is likely to construe a particular situation, it must be able to view
the situation from the other's point of view. In the Affective Reasoner
a robot's knowledge of other individuals' goals, standards and concerns
is built up by observing the reactions of those individuals to different
situations. This knowledge is stored in a database used by the Affective
Reasoner to enable a robot to predict the emotions of others, based upon
its experience of their earlier reactions in various situations.
Poker is an example of how, sometimes, what is said is not as impor-
tant as how it is said. Being able to infer meaning from how something
is said is not only an asset at the poker table, it is a sign of emotional
intelligence in general. Emotional cues that can be recognised by a ro-
bot, include not only characteristics of a human's speech patterns but also
eye contact and facial expressions, a furrowed brow, sweaty or trembling
hands, body language and skin colour (blushing). All of these forms
of emotional expression in humans can potentially be recognised by ro-
bots. Tools to aid recognition include cameras (to observe the eyes, facial
expressions and body language), microphones (to monitor changes in
voice) and bio-sensors (to measure changes in body temperature, skin
resistance and other physiological changes).
At the MIT Media Lab research into the recognition and measure-
ment of emotion has spawned a field of study called Wearable Comput-
ing, also known as Affective Wearables. A wearable computer is a wear-
able system equipped with sensors and tools that recognize its wearer's
psychological and other patterns, including any expressions of emotion:
a joyful smile, an angry gesture, a strained voice or a physiological change
such as an accelerated heart rate or a change in skin conductivity.
