In conjunction with National Engineers Week 2008, Harris Corporation will host a series of events at the two major campuses of its Government Communications Systems Division (GCSD) located in the
Melbourne/Palm Bay, Florida area. This year’s theme is "Engineers Make a World of Difference".
These events will feature demonstrations of various high-tech systems and products that Harris develops for
its government customers.
The College of Engineering and Computer Science of UCF has also being extended an invitation in hopes they find it worthwhile to demonstrate student projects, research areas, and curriculum.

On February 16, we will be hosting the SECME Regional Competition for young students in Seminole and Orange Counties. There will probably be up to 30 schools represented with 300-400 K12 students, parents and up to a 100 volunteers/judges from the student organizations and industry.
TeamUCF was invited to give a 45 minute presentation from 1:15 until 2:00 regarding our DARPA experience. This is something the young students will certainly relate to as it is engineering in practice. It will be a new type of audience of us, but it will be exciting to talk in front of so many people. Don is def

Knight Rider has been given the opportunity to participate in the annual "Otronicon" event in the Orlando Science Center. The Otronicon (www.otronicon.org) is basically a big 4-day gaming and simulation event with about 20,000 visitors from all over Orlando. That was a big chance for us to get some exposure to a multitude of visitors in all age classes. In the pretext of this event, Knight Rider got thoroughly cleaned and waxed (which is quite a rarity). Unfortunately, to allow the exhibition in the indoor areas of the Science Center, Knight Rider's gas tank had to be unmounted. Finally towed from the mechanic to the Science Center we had only a few inches space between the outside of the mirrors and the door frame while pushing it in. See some pictures of that in the Gallery.

There was excitement right up until the end in Victorville. CMU was scheduled to launch first, but they had some sort of problem with their GPS … which was attributed to the large Jumbotron sitting right next to them. Virginia Tech and Stanford went ahead of CMU, We were scheduled to launch fifth (having finished the qualifying event in fifth place out of eleven we later learned).

While it was a good run, our bot did not complete the 6 hour mission, going off course about 2 hours into the contest. We have since traced the problem to the GPS/INS. This is apparently a problem we have had since the beginning. One out of every 300,000 PS messages appears to be invalid, while the manufacturer's software indicates that it is valid. In this case they produced a NaN (not a number). Turns out this was the same problem that had occurred on course A a few days earlier that we had only partially diagnosed correctly. This NaN went in to the speed controller, which rejected it because of the fix we had implemented previously, but also got in to the steering controller and shut it down (it was not similarly protected). The car just drifted off course and into a driveway almost hitting a house.

We spent the morning testing the start process for the final event tomorrow. Things started off rough. Just as we were booting the computers on our bot, the generator cut out crashing everything. We scrambled to reboot. DARPA officials quickly showed up at our tent and asked where the heck we were. We managed to get to the start chute a couple of minutes late, but had to wait 2 hours before we did the first launch. The basic idea was that all 11 bots would be ready to go all at once and that they would be released in reasonably rapid succession.

UPenn was the first bot to be released but UPenn was e-stopped in the start zone (nothing they had done, but apparently some e-stop glitch). They restarted and completed a small loop just fine, but the next two bots never got anywhere. DARPA skipped these two bots, and went to MIT who got in the wrong lane on the loop road and could not get back to the start chute without going out much farther (we think they just wanted to go on a mapping mission). Stanford launched next and cut the first turn pretty tight, then immediately veered left coming within 20cm of Cornell and 50cm of our bot before it was e-stopped. It was not at all clear what caused that problem.

WE ARE IN! DARPA announced the finalists today in a ceremony at 11AM in the main tent. The first announcement that they made was that there would be only 11 teams in the finals. That was a far cry from the 20 teams they expected … and worse, 6 of the 11 were already known. The crowd was very quiet when this announcement was made. That left 5 slots for about 15 good teams, all of which were in the audience. These slots went (in order) to MIT, TeamUCF, AnnieWay, Honeywell/IVS and Oskosh which were added to CMU, Stanford, Ben Franklin (UPenn), Carolo, Cornell, and Virginia Tech. Of the 11 finalists, 7 had received the “Track A” award of $1M to build their car. Only four of us, AnnieWay, Carolo, UPenn and TeamUCF did not.

Unlike yesterday, today has been very low key (at least for us). We did one run on course C this morning, and we performed absolutely flawlessly, even meeting the mystery checkpoint. Still we will have to wait. We think we have done well enough to get in, but of course we could be wrong.

At the same time it was clear that a number of other teams were still trying hard to get in. CajunBot (the team from the University of Louisiana … go figure) attempted the long course B over and over until they finally smashed into an obstacle and pushed it into an abandoned house …. we suspect they are not going on. Team Lux also repeated the course, but ended up going down a one-way street … the wrong way. We suspect they are not in.

The team ran course B again … 6.5km with parking, obstacle avoidance, and some significant navigation. This was the run when we lost the sensor system the first day. CMU ran this same course again yesterday, and we heard they completed it in 21 min. That was 3 min slower than the day before so we were convinced that DARPA made the course more difficult … why I don’t know, since almost no one has made it through this course (we understand that only 6 teams have finished it with CMU being the only team to finish it twice).

Team Lux ran before us and did not get out of the start zone until their team interacted with the vehicle. The interactions cost them a boatload of time, and they had only 10 minutes to complete the course … they did not get very far. MIT went right before us and quickly got into the course and looked like they were doing great. After about 10 minutes we got word that MIT’s vehicle had gotten lost. They were driving around in circles going up and down streets in the little town looking for a way back out. The team had to retrieve the bot.

We spent the early part of last night trying to duplicate the conditions of test area A, but who has 12 counter-rotating cars and a large parking lot in the middle of the night? After a few hours we think we have something that might successfully complete one lap of the course … better than about 50% of the teams. At about 11 PM, however, I smell something … burning … in the car. That’s not good. Looks like we have fried the alternator. So its SUN night at 11pm and after a frantic Internet search we find a store 40 miles away open until midnight. Half the team heads there and the other half goes to the pit area. We get to the parts store just before closing, pick up the alternator, and manage to get back and install it by 3AM. Of course we need to run next morning at 9AM … and things are still not quite right.

We found the sensor issue and have fixed it! (we think ... damn Russian polygon clipping code). Its on to test area C today. We have learned a little more about test area C. In this areas we will have to navigate three 4-way stop intersections and 2 blocked roads to get through a series of waypoints. One of the roads will be blocked with barrels, the other with a large stop sign bar (like a railroad crossing bar with stop signs), held at eye level. Oh no … did I fail to mentioned we only look at things at car level, not eye level. Apparently some other teams have the same problem, and at least one team has crashed into the baracade losing a windshield in the process. We spend the morning re-calibrating one of our sensors to look at eyelevel, testing it with a series of safety jackets strung across a piece of string held by me. We only hope it will work.