In Good Hands
 

Hi,
you guys (and gals) are way passed me,
But it sounds Good,
start thinking how we can get our hands on some of the sys.
(read Sponsors)
and don't for get to check other posts, everything has to tie together!

rushed again gotta GO!
Geo.

Before we go into what we're going to use, how about we pull up a detailed map of the area? You should probably find something at the US Geological Survey site. I mean, yeah, it's going to be desert, but by desert, is the area the romantic-forest-gump-running-through-red-plains desert, or is it the Willie-E.-Coyote-falling-off-cliffs-left-and-right kinda desert? If it's the former, wouldn't radar be a bit overkill?

Ahem.. I think that even if it was a romp through the desert (which I doubt it is) that it would be nice to use Radar.. remember, this is a project sponsored by the military.. what technology do they use the most to detect stuff? :)

But it is a good idea though.

also, i don't remeber exactly, but won't the actual route not be released till the day of the competition? that way a team can't hard code in something, and it actually has to be autonomous.

that would mean looking at a map only tells you the terrain, which will help a little, but well, yeah, it can only tell you so much. plus, as randomperson said, the military uses RADAR in almost everything, why not use it in this? either that, or come up with something better, and i don't particularly feel like reinventing the wheel (although i'm sure it wouldn't hurt to try, if you really want to).

Better check radar out though. It may or may not work as someone all ready pointed out.
:edited:
I believe that it will not work due to something going on with the size of the wavelengths. Some obstacles may be too small to have the radar waves bounce off of. I think one meter is the limit.

hehe, i was the one that mentioned it may be a one meter resolution (i think) so yeah, it may not work for everything, but is there a better solution at the moment?

i'm thinking RADAR would be good for long distance stuff, detecting big rocks from far away, change course accordingly. then, find something with like, centimeter wavelengths (10^-2), and use that to detect close range stuff less than a meter. or just use it to detect everything, but it may come up with too much noise over a large distance (why a larger resolution may actually be better).

someone who knows more about RADAR and such really has to be contacted, cause this is all based on stuff that'd i've picked up from random things in life :). i think it's mostly right though. at least it makes sense, to me...

another alternative to radar might be cameras. I've been working on an experiment measuring the "growth" rate of objects as a variable of their size. while the data curve does show similarities to the graph of the function 1/x, i don't think my methods are up to par...
if my analysis is correct though, while a small object close up has a curve similar to that of a large object far away, a program could pick out objects based on the growth rate (as compared to other objects) and maybe use and use a range finding device (laser? sonar?) to verify the range, and from there, find the size.
just another (hair-brained? intriguing? confusing?) idea i thought i'd throw out here.....

CMUCam

Th CMUCam would be good for local object detection. If we mount a few pointing 360 around teh unit they could watch the area around it.

Pretty simple logic: Find what color dominates the most area, example. A lot of brown is the ground, a medium sized blob of grey is there, more brown means that grey is probably an obstacle, correct steering. Pretty simple design too.

don't know if you need 360 degree detection, but it definitely doesn't hurt. also, it might be better to use the CMUCam to look at where to point more sensitive instruments. it's an idea, don't know if it'll work, but something to consider.

could we possible use bluetooth in any way for systems interface?

using a range finding device plus cameras could work....