The big success? The food naturally. Etouffee.
Oh yeah, and Team Cajunbot did ok in the race too. Or so it says in their blog. (Dont’ you love living in a place where even gearheads and geeks have their priorities in order?) They didn’t get close to winning the 2 million grand prize but did get further than any team got last year.
Last year no team got further than about 7 miles into the “race” which pits autonomous, unguided robot-vehicles against a tortuous Mojave desert course. The contest is meant to push real-world application of cutting edge research–acting in the real world has turned out to be a pretty difficult problem. Computers don’t deal well with surprises and the real world if full of them. (Winning the race isn’t the point, in my judgment.) But this year a number of teams finished, mostly, it appears, by a combination of general hardware improvement, better software, and most notably the winners took their machines out into the Mojave desert for testing before the race. Experience, even if it is the experience of the programmers and engineers rather than the machines themselves, is important. Most interesting to me was that one of the top machines, from Standford, incorporated a learning algorithmn–the machine, theoretically anyway, was learning for itself. That is worth really watching. Learning, machine or human, is the really hard problem of both Artificial Intelligence and Human Education…it is to prod real-world applications of such outre understandings that the Grand Challange exists.
Anyway, just getting to the starting gate was a huge feather in UL’s cap. Bringing a rigged out swamp buggy to the Mojave takes a lot of nerve. Good for Team Cajunbot.