My work with smaller bipeds has been helpful, but in
many respects things are completely different with larger
models. Some things do not scale well. For example,
ants. If they were our size they could lift an elephant, but
they don’t scale well. (Except in cheesy movies). Biped
robots are the same.
With the small bipeds, I found it very easy to copy
the sequence of the left leg step and reverse it for the
right leg to complete a walk cycle. Small differences in
the mechanical structure didn’t factor that much in the
movement. Yet, that doesn’t work with larger structures.
We reversed the one-step sequence and found we
had to tweak almost every frame. In one part of the step,
we needed to add an intermediate frame to retain
It is the Sunday after the qualifications were due, but
we didn’t give up. WATSON took two full steps and
returned to the frame from which we started.
The DARPA Robotics
So, what of the challenge? Are we out of the
As it turns out, the short answer is ... yes and no.
This year’s DRC in December is mostly for Tracks A,
B, and C. Since DARPA is footing their bill, the folks
running the challenge mostly want to find out how well
some of their sponsored teams are performing. So, if we
were far enough along, we would have been invited to
test WATSON along with Atlas and some other terrific
works of engineering. However, Team Walk Like A Man is
still entered in the challenge that takes place in
Now that WATSON can walk, things will progress
even faster. We are still in the running for some of that
$2,000,000.00 prize money. SV
SERVO 12.2013 51
Be sure and view the video of WATSON walking
(available at the link article).