DARPA Grand Challenge

This is pathetic.

It was a contest to see which team could build an autonomous vehicle that would travel 250 miles across the CA/NV desert in the shortest time.  Over a hundred teams entered.  Most didn’t even make it past the initial qualification round.  Then fourteen were selected to actually attempt the race.

Not a single robot finished the race.  In fact, none got farther than a few miles past the starting line.  From the videos and photos, it looks like the teams spent thousands of dollars making their vehicles look like tanks, and then built the navigation and control systems out of Cracker Jack prizes.  One team spent three million dollars making their Hummer-based robot.

Every single one failed miserably.  Half of them didn’t even make it out of the gate, crashing into walls or going in circles immediately upon starting.

This is pathetic in two ways.  First of all, each team was given a CD containing GPS coordinates of the route.  All the vehicle had to do was follow the coordinates and not crash.  The one that went the farthest, the 3 million dollar Hummer, only made it 7.4 miles.  Now, I’ve never built anything like an automaton, so I am obviously not speaking from experience, but it seems to me that thousands or MILLIONS of dollars should be enough to make this happen.  Apparently not.

But what really bugs me about this -- what’s really pathetic -- is the fact that half of them crapped out instantly, and most within a mile or so.  This is pathetic because it means that all these people on all these teams spent all this money building these machines, and apparently did no testing whatsoever.  How can you spend thousands of dollars on a project and not TEST it once before the day of the race?  Because that seems to be the only logical explanation for why half of the robots -- make that half of the ones who made it to the second round of qualifications -- couldn’t even start the race, let alone compete in it.