Page 204 - Biomimetics : Biologically Inspired Technologies

Bar-Cohen : Biomimetics: Biologically Inspired Technologies DK3163_c006 Final Proof page 190 21.9.2005 2:56am

190 Biomimetics: Biologically Inspired Technologies

Sparked largely by the mid-1990s MIT graduate work of Cynthia Breazeal, sociable robots
integrate many of these human-emulation technologies into singular synthetic organisms, designed
to communicate more effectively with people (Breazeal, 2002). While these robots only crudely
simulate social cognition, they are being actively used as modeling tools in cognitive science (Fong
et al., 2003). Since Breazeal’s seminal work, a sizable number of sociable robots have sprung into
existence. Although a comprehensive list is beyond the scope of this paper, a few sociable robots
include: Ridley at MIT (lead by Deb Roy), Nursebot Pearl of CMU, Kismet and Leonardo of MIT,
and Mabel at the University of Rochester (built by a team of undergraduates). Additionally,
companies including Panasonic, Sony, and Honda have lately pursued sociable humanoid robots.
Although these robots all seek to achieve bio-inspired communicative interaction with humans,
none has a realistic humanlike face.
In social robots that do have faces, the hardware mimics the expressive action of the human face
— humanity’s primary mode of expressing affective states (Ekman, 1989). Whether depicting a
realistic human or an abstract character (like a cartoon or animal), the expressive animated motions
of the character should be humanlike in order to be sensible to a human, because the human nervous
system is innately and finely attuned to understand the human face’s visual language (Levenson
et al., 1990; Bruce et al., 2002).
As discussed later in the chapter, this can be a challenging hardware task, and even harder can be
the socially interactive use of facial expressions. Better mechanization and automation of this social
expression could unlock many useful service and entertainment applications from toys to comfort-
ing companions for the elderly. Even in a military scenario, wherein a robot must communicate
swiftly with human soldiers, the power of emotive communications cannot be over-estimated.
The mechanics of the biological human face are well studied; meanwhile the semiotics of
human-facial communication have been preliminarily defined by anthropologist P. Ekman and
others in the aforementioned FACS (Ekman and Friesen, 1971). Body language, also well studied
(Birdwhistle, 1970), can further enable robotics’ sociable applications. While further work remains
to decipher the cognitive systems that underlie dynamic facial effect (including their complex
relation to language), these robots can still be interesting as entertainment, training devices, and as
quantitative tools for the better study of social cognition.
Most hardware technology for simulating facial expression springs from the entertainment
special effects industry, where the technology is used to animate characters in movies, theme
parks, etc. Stan Winston Studios, Walt Disney Imagineering, and Jim Henson Creature Shop, and
many other ‘‘animatronics’’ (themed animation robots) shops, utilize the power of nonverbal
communication, by simulating human and animal faces and figures in story-telling context. In
these applications, the complexity of social cognition is theatrically designed by animator and
writers, and is not interactive or intelligent. Nevertheless, these approaches that emphasize com-
mercially presentable results have achieved the highest degree of mechanical aesthetic biomimetics
in history (see Figure 6.9), as shown in feature films such as AI and Jurrassic Park among others.
Animatronics seems like a natural match for sociable robotics. Indeed, the merger of anima-
tronics and sociable robotics has begun; in 2002, one of the leading shops in animatronics built the
mechanical and aesthetic systems of Cynthia Breazeal’s Leonardo robot (Bar-Cohen and Breazeal,
2003; Landon, 2003) (see Figure 6.10).
As with the work of Luc Steels and Qrio, the Leonardo project uses learning algorithms with
vision–tactile–language knowledge fusion to accomplish learning by imitation. This work is
collinear to that of other MIT groups such as the Cognitive Machines Group (CGM) led by Deb
Roy, in which the robot Ripley uses a grasping mouth to manipulate objects. Ripley clearly has a
machine identity, and no facial expressions. Leonardo, by contrast, boasts 32 DOF in the face,
achieving very agile facial effect. While Leonardo is anthropomorphic, Leonardo is conscientiously
not human in form (Breazeal, 2002). Leonardo team leader Cynthia Breazeal expresses that realistic
animatronic technology is not quite human enough to be convincing, and just human enough to
push people’s expectations of the intelligence of the machine and too high to be met with today’s