Wednesday, August 7, 2013

Update on Past two days -- Rough Preliminary Version

Okay all, we have let you down again on the blog updates. It is not completely our faults, I promise, there were severe technical difficulties in the way; we ran into problems with blogging itself, but also with our vehicle. So if nothing else, the fact that we haven't updated should tell you all that we at least have our priorities straight, because we were busy fixing things.

Before we get too far into this, we would like to share the presentation that was mentioned earlier.  You can find the technical presentation that was delivered by José Gomez here: https://www.youtube.com/watch?v=yB4lDshla4s

A LOT has happened since the last blog update. We want to start from where we left off, which was late last night (aka, the night of August 6th and into the morning of August 7th). The easiest way to explain it is that EVERYTHING BROKE AND WE HAD NO IDEA WHY. Honestly, everything, it was overwhelming. From software to hardware the aspects of our vehicle were picked off one by one as if they were the stars of some kind of bizarre horror movie.

This put the team into a bit of a scramble. We basically had to pull together for a divide and conquer approach and just try to tackle everything at once. Things started reallllly going poorly around 11 pm/12 am. We experienced a few severe crashes while simulating end-to-end missions of the IARC, and sensors started acting funny; the use of the word funny here is very loose, because the behavior was erratic and incredibly frustrating to deal with.

By 1 am we were unable to do any good flight testing. Our vehicles were unstable and flying with bugs that seemed impossible to find and fix. However, we knew that the competition flights didn't begin until 10 am, so we were set to work through the night up until the competition began.

Around 6 am things were calming down. A lot of problems had been cleared out, and we could begin flight testing again, but we still had a ways to go. A lot of the team had succumbed to naps in various places that weren't intended for napping, but we always had at least five to ten people awake and working on flight testing and vehicle repairs.

It was at 8 am that we had to make some strategic decisions. For competition each team is given four trial runs and one pass. We knew that we had a lot of work to do before we had a chance and completing the mission, so we sent eleven of our members to represent MAAV at the competition site, while leaving four behind for testing. The rationale behind this was that it was impossible to test once we were at the competition site, and the vehicle at 8 am was unable to complete the mission. So the goal was to have eleven members present to offer a pass on our first turn, and then to continue stalling until our test vehicle was ready to fly.

And so we went forth with our plan, with time ticking and the stress really weighing on the entire team. We lucked out a bit in that we ended up drawing last in the flight rotation, which gave us the most possible time to maximize our system. We still weren't ready for our first turn, though, so our eleven members on site went to stalling and officially passed for our first turn.

Meanwhile, things were being cleaned up on the competition vehicle off-site, but still were not looking as clean as we had hoped. Knowing that our first flight time was fast approaching, the team made the decision to cut our losses and show up at competition with our not-so-perfectly tuned vehicle. We figured that we had a better chance of completing the mission with a vehicle that contained "mystery bugs" than we did of solving those mystery bugs and developing a perfect vehicle in time for our flights.

So we showed up in time to compete for all four of our allotted trials. During our first two runs our vehicle experienced an extremely bizarre failure where it would kill mid-air without any rhyme, reason, or warning. The behavior was terrifying, and we were worried that we wouldn't even be able to get a chance and recovering the flash drive. Fortunately, we were able to fix it and were autonomously exploring the course by our third attempt. During this attempt we autonomously explored the compound, located a table, and then approached the table to search for the flash drive. Unfortunately, it was that last absolutely essential step that failed, and even though the vehicle "swept" the table it was unable to find the flash drive due to some turning errors.

Based on our third run we tuned our vehicle, changed parameters in a way that we thought would maximize our chances of success, and then went forward into the unknown. We must unfortunately announce to you, though, that our fruits were unsuccessful in that we were unable to fully complete the mission statement.

I know that this is a lot of information at once, but we wanted to let you all know exactly what our status was. However, I, the current author of the blog, have been up for 48+ hours straight and am exhausted. We have a lot of videos and pictures to share, and tomorrow I will be elaborating upon what exactly happened at competition. Apologies for any typos, but this is intended to be a rough update for our situation.

1 comment:

  1. Through every failure comes success!! Make sure you get plenty of rest before hitting the road!

    A concerned citizen. :)

    ReplyDelete