Results
The results of this investigation turned out to be
different from what we expected. Robots did not walk the first time we
built them. When they did, their legs were not responding at the same
time and they lacked traction. As we modified the connection scheme to
make the robots walk in different sequences, some of the Nitinol caught
fire. Additional Nitinol had to be requested for repair. Even though we
had already predicted that the original current would be too big for the
modifications made, the resistances used to prevent this did not work
for us as we had planned. We also had trouble finding an appropriate
walking surface for the robots because it had to be porous and at the
same time smooth enough to let the legs slide on the recovery part of
the sequence. At the end, additional weight (two pennies) had to be
installed on the robot to provide for proper traction on the surface
that we selected.
We found that the most efficient walking sequence is
the TRIPOD. It delivers the maximum current-force relationship. The
walking sequence was the most stable. It walked in a straight line all
the time. It had a medium velocity and an equal pulling strength.
Another interesting fact that we discovered was that the legs that were
not being moved, lifted a little off the ground while the others were
taking a step. This happened because the Nitinol seems to pull the legs
back but also down. This is good because it provides tremendous traction
on the active legs but on the other hand, it limits the traction of the
static legs. This does not give the static legs the stability they need
to hold in place once the active legs are relaxing. This was fixed by
readjusting the crimps.
We will make a quick comparison of the walking
sequences. The TWOPOD was supposed to walk in a straight line and to
have the maximum speed, but it danced its way across the finish line. It
started to walk crooked and the process of applying the pulses at the
correct time, one after the other, was troublesome since three
connections had to be made. The controller was also a big problem. It is
not appropriate for the comparisons we were making since it is not easy
to handle. The Caterpillar sequence performed well but it also turned
instead of walking in a straight line. This is primarily because the
Nitinol on all six legs does not respond at the same time, even though
proper care was taken to apply equal tension to all the wires. We found
that it is most likely that future robots constructed will need
additional weight since the body itself is not that heavy and does not
provide good traction (all of this refers to the walking surface we are
using and that have previously mentioned). Once all the measurements
were taken, the TRIPOD was the most efficient.
