The final standings from the
competition. We finished in second place.
The final points
break down from the competition.
The Rules for the competition.
Our Journal Paper for 1999.
With a little more time (or just better planning on our part) we could have
gotten all six gates realiably.
|
...T-minus three weeks
We needed to get pool time in well before the competition if we wanted
to have a chance to win. We had set a deadline of one month, and we
almost hit that.
We were able to acquire two hours of pool time four days a week,
and it really helped. If we were more efficient with our time, we
probably could have gotten even more done.
|
...T-minus two weeks
|
...T-minus one week
...add info about change the sonar depth finder to get better
resolution in the shallow waters.
|
...T-minus
|
...T-minus 2 days
We still haven't tested the isobath software, and the only
real place to try it was at the pond.
The code for the sonar gate detection was in place, and seemed
to be working in our pool. However the pond at the competition
would be very much different, and if we wanted to tune it for
that environment, we would need lots of pond time, and 'hopefully'
good images.
The prop shrouds were in place, but they were too flimsy to
be of any benefit. Using a large hose-clamp and bolts, the
shrouds were firmed up enough so that you could almost lift
the sub by them. However this took up most of the day. We were
able to rush over to the pond, put the sub in the water and take
some images from the sonar to tune the gate detection routine.
Taking all the equipment back to the room and looking over
the gate data, we find that it's to difficult to distinguish
the gate from the background. The clutter in the pool, and
the pool's overall shape, make it hard to extract the gate.
So all the data taken is just about useless.
(Mental node to self, make sure we can view the data on site
to be sure what we're taking is usefull.)
|
...T-minus 1 day - Static Judging
We arrived in the morning, and set up our tent, and our hardware
to display our user-interface and the screen shots from the sonar.
After the static judging, we were ready put the sub into the water
and run it around. Because of the amount of teams that needed to
use the pool, we were initially restricted to an hour of time, with
the first half-hour being able to start where we wanted, and the
second half would be based on the incomming teams preference.
Since we were up, we chose to tune the sub to head for the first
gate (a logical choice). We had put the wireless ethernet antenna
on a cylinder of styrofoam so that it would remain afloat while the
sub ran at depth. That way, we would have complete communication as
it traveled around the pond. Giving us feedback, and access to change
parameters as we saw fit. (As it turned out, there were some spots
on the far end of the pond where we lost communication).
We had two goals for this session, tune the sub, and try to get good
images from the sonar.
|
...T-minus 6 hours
We were the first to have pool time on the day of the competition,
and arrived early to take advantage if it. We needed it. Scott had
made some changed to the isobath following, and they needed to be
tested. Because of the ominous looking weather, everyone's hour
of pool time was cut to 40 minutes.
The sub was lowered, and after a few false starts and a few parameter
that were tweaked, like magic, it traversed the course and made all six
gates. Finally! (let me say that again, finally, Finally,
FINALLY! ) Our diver was sooooo suprised, and commented,
"Wow, I guess you guys didn't go to Clubing last night!"
However, this was with the sub trailing the wireless ethernet antenna.
We needed to try with the antenna in the stowed position.
Now with the antenna stowed, we set the sub off again around the
course, and it would hopefully clear all six gates again. But,
since the dynamics of the sub had changed (less drag, now that the
antenna was stowed), the sub oscilated more and we missed gates.
The weather had gotten worse (a rather ominous start to the competition),
and we were all told to leave the base since there were lighting
strikes close by. Our pool time was up, and we needed to leave. We
had a good run, and would have to do some untested tuning to smooth
it out.
Since we had an hour to kill, we decided to get some breakfast and
just hand. So that's what we did.
|
...T-minus 3 hours
After the hour, the weather still hadden't cleared up, but we
headed back to the pond to see what the status was, and how things
were going to be handled.
We should come back in two hours.
|
...T-minus 1 hour
Time passes...
With a break in the weather, they decide to try to squeeze in as many
of the contestants as the weather would allow. Since there was nothing
for us to do, we could actually watch the other competitors. That was a
big change from last year.
First up was FAU. They had the cutest sub, not to mention the coolest.
|
...T-minus 30 minutes
Stevens was next.
|
...T-minus 0 seconds
So now it was our turn. We had ten minutes to perpare, but
didn't really need it. The wind had picked up since our
practice run in the morning, which caused a good chop on the
surface of the water. This had us concerned.
The interface was brought up, and we check that we had communication
with the compass (a plus over last year) and that the pressure sensor was
working properly. The motors were spun to be sure they were operating
correctly also. All was good - hardware wise. Now for the true test of
the software.
|
...T-plus 5 minutes - 20 minutes of competition time left
Now starts our trial for the competition. The code was run, and
the kill switch activated (so that the motors wouldn't spin). Down,
into the water, the sub was lowered, and the kill switch was
de-activated.
The sub quickly dove to the programmed depth and tried to track
the isobath. However, either due to the untuned code, or the strong
chop in the water (or both), the sub had difficulty tracking
the isobath. Even though it was doing the same things as before,
it seemed like now it didn't have enough power to initiate a correction.
Just to be certain, the sub was stopped and pulled back to the start.
Sure enough, it looked liked it was trying to make the correction,
it just didn't have enough power to do so.
|
...T-plus 8 minutes - 17 minutes of competition time left
We pulled the sub out of the water to adjust the downward looking
sonar (tilted up to look slightly more forward), and upped the power to
compensate for the drift, and give it more power to play with.
The code was recompiled, the interface was brought up, and everything
checked out. Back into the water it goes.
(Mental note to self, take ALL pertinent parameters from a text
file to reduce the amount of compile time necessary.)
|
...T-plus 9 minutes - 16 minutes of competition time left
Again it was set of on it's merry little journey. This time
it was able to make the correction and veer away from the shore.
Alrighty, let see what it can do! We needed at least three to
move into first place.
There still was an oscillation in the tracking of the isobath,
but our strategy was to let the sub circle the pond until it
either passed through all six gates, or we ran out of time (hoping
for the former).
The sub made it throught the first and second gate, now for the next
challenge, clearing the tight left-hand turn between the second
and third gates.
While I think the sub just about scraped by, it did make the turn,
however this set up some ocsillation that didn't get a change to dampen
throughtout the rest of the run.
Despite our code, the sub was determined to make it through the
third gate. It took a little turn out toward the middle of the
pond, but then decided to come back and through the gate.
There was another somewhat tight left-hand turn between the forth
and fifth gates. We also made that turn, however, we fell out of
sync for the fifth and sixth gates. We passed on the outside of the
fifth gate (doh!), and on the inside of the sixth gate
(doh doh!)
|
...T-plus 18 minutes - 7 minutes of competition time left
Since we circumnavigated the course in about 8 minutes, we really didn't
have enough time to go around again. And since they were trying to
finish off the competition as soon as possible, we finished our run
with four gates.
Needless to say, we were very pleased with the performance of the
sub. Bad weather? Not a problem, it was a bright day afterall.
|
|
