The Greatest
Playground of All
by John R. Sosoka
The real world, what a place! The
smell of freshly mowed grass, the
breeze against my face, textures,
colors, shapes. The real world is
incredibly compelling to us. Life, from
plants to animals to other humans,
has an incredible draw for us. We’ve
been finely optimized to survive in the
real world. Looking at a beautifully
rendered apple on a screen is a treat
for me, having been a part of the
computer graphics community for
longer than I care to remember. But
holding, touching, and biting into
the apple in my hand is an entirely
different experience. Primal and
glorious, a feast for my senses.
Though I started out in a machine
shop, much of my work life has been
spent in the virtual world. The journey
back to the physical world has been
very interesting. This, I think, is one of
the attractions of robotics. Rodney
Brooks showed us how to get robots
out of their heads and into their
bodies. The real world became the
model; reactions were faster and
subsumption provided an inkling of
the robustness that we see in nature.
It was hard not to see cockroaches in
a whole new light.
When we strive to create robots
that can exist in the same messy,
ever-changing, complicated world that
we live in, we are faced with
tremendous challenges. And through
the process of solving these
76 SERVO 08.2008
challenges we get back something
very precious in return. We get to
experience awe. In some ways, it is
the childlike awe of a thunderstorm
but now we have ‘knowing’ mixed in.
Last summer, I was working on a
vision problem related to navigation.
A tremendous amount of processing
power is required to deal with the
per pixel optic flow algorithms that I
was exploring. Then a fly swept in,
neatly avoiding my hand and landing
on the edge of my cup — all this
in a breeze.
Before my robotics adventures,
that fly was just a nuisance to be
eliminated. Now I sit amazed by this
self-contained, autonomous creature.
It runs on garbage, procreates,
navigates using optic flow, and
aerobatically escapes predators.
Understanding how hard these things
are to do, I’m left in a state of
informed but profound awe.
A diverse collection of folks have
gathered at Ugobe to tap into bits of
this inspiring experience of ‘creating’
life. Part of the fun of robotics is that
you get to work with people with
vastly different skills. David Calkins
captured this perfectly in his column
(SERVO, June ‘08), “What the Heck is
a Robot, Anyway?” The need for us
to stretch to understand each other
is hard, but once you get through it
the teams are incredible, eclectic,
and electric.
Just to give you a taste of this,
here are a few of the crafts and
disciplines to be found at Ugobe:
illustrators, sculptors, animators,
model makers, CAD operators,
firmware engineers, electronic
engineers, computer scientists,
performance artists, ethologists,
cognitive scientists, synthetic
biologists, voice talent, recording
engineers, mechanical engineers,
writers, testers, spouses, children,
cats, and dogs.
Life Forms
At Ugobe, we are working toward
robotic technology that can capture
some of the key elements of life at
the level of a non-human animal.
While I’m a great admirer of David
Hanson’s work on conversational
humanoid robots, we’ve initially
set our sights a bit lower on the
evolutionary ladder. From a technical
standpoint, even creating a realistic
robotic mouse is beyond the current
state of the art. With our life forms,
we are working to advance the level
of realism and autonomy of our
creatures while still making them
accessible to a wide audience.
Human-robot interaction is a
very active field right now. Many
projects in this area are examining
the role of affect and the underlying
emotion models. One of the problems
