Letter Last year, we had the pleasure of meeting up with Warren Williams and Bill Zimmerly - better known as Team Phantasm. This two man team, based in St. Louis, made it to the end stages of the DARPA Grand Challenge before pulling out at the last minute due to a series of technical difficulties.
Williams and Zimmerly represent a lot of what was right about the $1 million Grand Challenge event. With a meager budget, they came up with an innovative vehicle design and poured an incredible amount of energy into their project. Their failure, if it can be called that, to compete in the end also shows just how hard building a robot vehicle than can cruise hundreds of miles really is.
It's a shame that all the vehicles struggled and that no team managed to go more than 7 miles. But it's safe to say the robots will be back in action again in the near future. And to give a picture of what this contest meant to Team Phantasm, here is a letter from Zimmerly that seems to sum up the event pretty well.
Dear Friends who have covered our adventure,
This was easily one of the highlight events of my life, and I think DARPA did a great job of coordinating this difficult, first-of-it's-kind event. I especially enjoyed the fellowship among the teams with many of them sharing parts, tools, and even food with us.
Warren told me that the Robot War events are just like this, where even teams that are just about to do combat will often provide the other team with parts, tools, and even expertise to help out the other team. I struck up many friendships with the teams and even got to meet the legend himself, William "Red" Whittaker of CMU.
Our trip to California was uneventful and went quickly, but when we got here, we had nothing but trouble. The radar system was working fine in St. Louis, but it wasn't working when we tried our first test at Camp Pendleton, and we weren't able to fix it in spite of dozens of calls to the manufacturer, Epsilon Lambda.
We received an email from a group of Buddhist Monks, asking us if they can come and "Bless" our vehicle. Being a Christian myself, I would have preferred that they not do this, but I'm not the team leader, and Warren thought that it would be a great media-attracting story, so I reluctantly agreed to give them a ride from the registration area to the garages. Warren was right -- your brothers and sisters in the media flocked to our section to witness the "blessing".
We spent most of the evening completing the sonar transducer and had it working quite well by bedtime of day one. Well enough, that we decided to use sonar only because of our radar outage. It seemed as though it would work so long as we stayed under 10 MPH.
But the Radar was just the beginning of our problems, as the next one was a far more devastating one: a power surge ran through the building where we were working on day two and it fried both the motherboard and the power supply on the Ladibug. This prevented us from making a QID (qualification) run the third day, and had it not been for the Kawasaki reps paying us for a new motherboard, we would have had to quit then.
On Wednesday morning, our first QID run was a failure because there wasn't enough power to run the computer and it kept rebooting.
Another similar problem was caused by the flakiness of the Hard Drive. When the computer was running and Warren would do something as simple as lifting the self-righting mechanism, the computer would crash. He said that it was because the Hard Disk was just too sensitive.
We tried to load the OS and Ladibug software on a Flash Disk, but were unable to get it to boot. (It was too tiny to load Windows 2000 on it directly, and we had no way of writing a Master Boot Record, aka "MBR", to it any other way.) We finally spent huge amounts of time loading Windows 2000 and all of the associated device drivers to a laptop hard disk -- but after doing this, we had Wireless and Serial Comm problems. Arg!
The power problem finally did us in however, since we weren't able to solve it. Each and every time we went to the starting gate of the QID, and started up the computer, it wouldn't stay "up".
But all was not lost as everyone just loved the design and look of the Ladibug. From my perspective as a programmer, my work can't be observed in the physical world like Warren's work on the vehicle itself, so I was disappointed. As it turned out, my software never got tested, since we could never keep the computer running long enough for it to sense and respond in an autonomous way.
I'm very much looking forward to the next running of the event -- sometime in 2005 or 2006 -- but unfortunately it looks like Warren and his wife had had enough. Unless they can get some major sponsor to pony up the money for DGC II, they probably won't return.
I however, do plan on a return to the event. If not with the Ladibug and Warren, then with something that my brother David will build. At least we'll have 18 to 24 months to properly prepare for (and TEST!) instead of the paltry 3 and 1/2 months that we had for this event. (Note: Warren began stripping the ATV and preparing it for the December 1st site inspection on November 18th.)
I've had a taste of the great event and I want more! It's "in my blood" now and I fully expect to get better sponsorship for my next appearance in DGC II.
- Bill Zimmerly