Control Engineering

Link This | Email this | Blog This | Comments (0)

Team #29 Natalythe Engineering
September 26, 2007

Team #29 on my list is Natalythe Engineering (no website) from San Diego, CA.  N E did not get one of the million dollar development grants nor were not selected to be on the list of 36 teams going to the National Qualification Event (NQE, the Oct. elimination trials) which puts them in the same situation as the other 53 teams not selected.  The question then becomes do you keep working to solve the problem or do you give up and walk away from the project?  Do you shift your focus from winning the race to winning market share by commercializing your design?  These are tough questions each of the 53 teams has to answer for themselves and might not be ready to answer in front of the control engineers of the world just yet.

Here is a picture of the N E crew in San Diego, CA.



Their vehicle Quondepo is shown below spun sideways in the road for the picture - a very Starsky and Hutch like maneuver.  Note the dent in the drivers door panel.  This was caused by making a left turn past a parking lot lamp post concrete base after the nose sensors had passed it but the rest of the car had not.  The N E team made the necessary corrections to the software shortly there after.  Watch for this characteristic dent on many of the other teams vehicles.   This is a lesson in robotics that each team has had to learn and most learned it first hand.



The German version of the Wikipedia entry for the DARPA Urban Challenge has maintained the original list of teams and highlights them as their status changes rather than dropping them from the list as in the English version.
http://de.wikipedia.org/wiki/DARPA_Urban_Challenge_2007  If you don't read German, run the page through http://babelfish.altavista.com/babelfish/tr to see it in English.  There are 89 teams for this race, many are still working on solving the problem of how to make a vehicle driverless in spite of the fact that DARPA has said that they are no longer eligible for the $3.5 million in prizes.  Only three out of the 89 were going to get prize money so we all knew that 86 teams were not going to get any (odds of 30:1 against your team). When do you quit working?  If you were working thinking you were going to be one of the eleven golden teams to get a million dollar development grant do you quit when you don' get it or if you were working thinking that you were going to be one of the three winners of the Final Event (FE) do you quit when DARPA tells you the bad news?  Luckily most every person on every team is working on solving the problem because it is an interesting and challenging problem.  Prize money is way down the road, long after you have solved the problem.

Of course if you throw enough money at this problem you can solve it using very expensive parts.  Some would say that a supply truck can only drive itself if it is outfitted with flight certified missile parts and software.  You have seen cruise missiles on TV arriving at the front door of a particular street address.  Cruise missiles cost US tax payers $1.5 million dollars each.   So the problem is not can we outfit a cargo truck to go to a specific street address and wait to be unloaded - because we can already do it for $1.5 million dollars a copy. The question is can we do it for an affordable price.  What is an affordable price?  Well we are talking about an accessory for a military truck.  How about 10% of the trucks cost for the automatic driving feature - that sounds reasonable.  Military cargo trucks cost about $380,000 each - a bit more than their civilian counterparts because of features and materials that the military has asked for - 10% of that would give you $38,000 to work with for the production version of the guidance system.  People best at finding the cheap way to get the job done are people who have no money - that would be most of the garage companies around the country who have been formed to be in this race.

Economy in the extreme.
While looking for reliable yet inexpensive switches for AIM's demonstration vehicle AGV WENDY DARLING normal rocker stitches from a normal supplier were quoted at $68 each. I  kept looking and eventually found a SPST switch for $0.97 that is not very pretty but works.  The SPDT version is $1, the DPDT version of this switch is $2.  While looking for panel fronts to mount the switches on, I found a preformed one that holds them in groups of four for $1.55 - not space flight certified but this is after all a prototype we are building to experiment with.  The prototype needs more switches than the production version to allow us to rapidly change the configuration while testing.  Finding the least expensive switch does two things, it identifies a less expensive alternative switch that might not otherwise been found and makes our budget go farther.

GO ROBOTS !

Posted by Paul Grayson on September 26, 2007 | Comments (0)