PDA

View Full Version : Cornell Minesweeper Pictures!



Ironman
02-04-2009, 08:04 PM
Well here is the chassis of our robot ladies and gentlemen.

Feedback is always welcome.

most of the sensors are missing, but they should be up within a week, i'll post more pictures then.

i call him squeaky because his wheels sometimes rub up against the aluminum.

the picture of the hub i had to take because it's a custom made bicycle hub for our axles, it's some intense work right there. I think each one took 6 hrs of machining.

elios
02-04-2009, 09:43 PM
wow. idk what to say....

robologist
02-05-2009, 02:55 PM
Pretty cool. What was the rational for angles wheels? Any links to the hardware pictured? Motors, h-bridges, etc?

Ironman
02-06-2009, 12:49 AM
no rational.

originally it was to give added stability.

it does give a wider base, but not a significant amount to actually say it's more stable.

metaform3d
02-06-2009, 03:17 AM
it does give a wider base, but not a significant amount to actually say it's more stable.I've seen this design on a number of bots. So the question is: what's the optimal angle for a given weight? If the bot is very light then angling the wheels like that increases the wheelbase, but if the bot is heavy then tilting the wheels puts added strain on the axles. What are the tradeoffs?

Also, is that a brick on the back end? What, it wasn't heavy enough?

lnxfergy
02-06-2009, 08:41 AM
There is one other optimization beyond just widening the wheelbase: if you are cornering at high speed, the wheels are better setup to handle that force.

-Fergs

Adrenalynn
02-06-2009, 01:01 PM
And will handle better [theoretically] at high speeds, sharper cornering [see: toe-in/toe-out in auto-racing] Although that's an awful lot of toe-out...

Connor
02-06-2009, 03:55 PM
And will handle better [theoretically] at high speeds, sharper cornering [see: toe-in/toe-out in auto-racing] Although that's an awful lot of toe-out...

I believe it's Camber, not toe-in/toe-out. Toe-in/Toe-out would be the other axis.

http://en.wikipedia.org/wiki/Camber_angle

Thanks, Connor.

No0bert
02-06-2009, 04:40 PM
Dont you use more battery power though?

Because now that I think of it, the center of gravity is shifted to some place other than the center, and so the motors are expelling more torque than required to keep the bot moving and stable, so because of this added torque, wouldn't more current be flowing, using unneeded amperage?

Or am I speaking gibberish?

Rudolph
02-06-2009, 04:43 PM
Also note the significant negative camber of racing wheelchairs (http://images.google.com/images?q=racing+wheelchair&btnG=Search+Images&gbv=2), which was the first thing I thought of when I saw Ironman's photos :)

robologist
02-06-2009, 05:17 PM
Dont you use more battery power though?

Because now that I think of it, the center of gravity is shifted to some place other than the center, and so the motors are expelling more torque than required to keep the bot moving and stable, so because of this added torque, wouldn't more current be flowing, using unneeded amperage?

Or am I speaking gibberish?

Maybe a little gibberish. :)

I think the "splayed" design is somewhat more stable efficient in getting over obstacles, lowering the center of gravity some, and providing more a push inward towards teh carried mass instead of upward on a bump. Somewhat like the sweeping of a bug leg as it propels forward, cf Robert Full at the Polypedal lab, Berkely(sp).

I'm still intersted to learn what motors and h-bridges had been chosen in this build up. Little details like this sometime get glossed over, and it would be a really good data point to know.

lnxfergy
02-06-2009, 06:02 PM
Also note the significant negative camber of racing wheelchairs (http://images.google.com/images?q=racing+wheelchair&btnG=Search+Images&gbv=2), which was the first thing I thought of when I saw Ironman's photos :)

We discussed this in the chat the other night, it seems this is actually for the purpose of racers not hitting their arms, not for traction.

-Fergs

Adrenalynn
02-06-2009, 06:58 PM
I believe it's Camber, not toe-in/toe-out. Toe-in/Toe-out would be the other axis.

http://en.wikipedia.org/wiki/Camber_angle

Thanks, Connor.

dang it. Absolutely right. I was running out the door as I wrote it, my ride was already waiting, and I totally misspoke. Mea Culpa.

Ironman
02-07-2009, 12:10 AM
the brick is a simulated payload.

actually the robot isn't heavy enough.

the center of gravity is 1/10th of an inch in front of the axle.

we do run off of battery power, but we like to have wall power to test.

the motors its using are brushless motors from anaheim motors and they come with a motor driver.

they're nice clean package. We didn't make the h-bridges.

if you want more specifics about the motors i can post them, just ask.

robologist
02-07-2009, 09:23 AM
Yes, some more specifics about the locomotion system would be helpful The model number of the motor, what type of gearbox it has, the gearbox model number, the controller model number all would be good. As would any encoder info or hall effect sensor info.

Ironman
02-07-2009, 03:40 PM
Gearmotor Anaheim Automation BLYSG342S-24V-3000-R75 Motor Driver Anaheim Automation MDC151-050301
the last part is the model number.

motor
http://www.anaheimautomation.com/brushless-dc-spur-gear-motor.aspx

motor driver.
http://www.anaheimautomation.com/brushless-dc-driver-mdc151-050301.aspx

robologist
02-07-2009, 04:08 PM
Thanks, gives some interesting specs to look over.

Ironman
03-01-2009, 01:06 AM
Alright so we got the 'bot anodized. All of the sensors are now mounted on the robot. But I fried the camera two days ago when I was testing a fire wire cable. Ya don't test with stuff you're actually going to use. Never quite the best idea.

The back of the robot has been moved back 6 inches to move the COG (center of gravity) behind the axle. Initially it was 1/10th of an inch in front of the axle, and now it's 1 inch behind the axle.

The coffee pot looking thing in the front of the robot is a LIDAR (laser radar) made by SICK. If you take a close look at the second picture you can see the GPS antenna. It feeds into an INS (inertial navigation system) that gives an absolute position accuracy of ~1cm. The INS is made by Microbotics inc. They're a pretty cool outfit, great if you have any wish in flying robots. lots of nifty electronics you'll need.

The drivetrain now has an addition of an electronics protection system of it's own (it also acts as a nice handle). The drivetrains on each side are completely modular and can be removed in under a minute...

The main electronics box is rainproof and dustproof (IP65 Rated ). It's also on drawer slides so that it can be easly accessed.

we have a garage door opener that acts as our wireless emergency stop button, and an Intel core 2 duo with 2 gigs of ram for processing power.

The onboard emergency stop button is mandated by our competition and does exactly what it says it does.

Our battery is a 51.8V battery that gets stepped down to 24V. Everything else has it's own power supply that can step down 24V or it runs at 24V.

Questions? Critiques? we're always looking to get better.

droidcommander
03-04-2009, 03:12 PM
<geek envy>

Ahhhhhhhhhhhhhhhh

robotty goodness
</geek envy>

DroidCommander

Ironman
03-10-2009, 06:01 PM
The robot's official new name is: Nero.

Ironman
07-29-2010, 02:36 PM
Hey Guys,

It's been a while since I've been on Trossen, but the robots keep rolling.

Here are some images of our latest robot: Aki-ra (named after this dude in Cambodia who runs around with a stick demining)


http://hphotos-snc3.fbcdn.net/hs478.snc3/26199_396105325578_503600578_4951825_572235_n.jpg

You'll notice that in one of the images the robot is set up to run a test on some inert landmines.

Also, the nuclear quadrupole resonance was never deployed mostly due to the supplier never answering our phone calls.

Unfortunately this will be the last robot developed by Minesweeper as our funding was cut due to budgetary restrictions at cornell.

Stobs
09-13-2010, 10:38 PM
Very cool project, from both an engineering prospective as well as it's intended end-use - far too many global spots were land mines have been buried and "forgotten." I know this is going to hurt - me! lol - but what kind of a materials budget did this project eat up?

Ironman
09-14-2010, 01:39 AM
The last project cost about 7k to put together. Of course about 4 to 5 k was the metal detector. Unfortunately the project was canceled by Cornell due to funding cuts.

Stobs
09-14-2010, 10:44 PM
Thanks for the info Ironman - that's a lot less than I had thought it would be, so the amount of work you and your team put into that is very impressive! Very sorry to hear about the funding cut, which was no doubt a disappointment.

Possibly, if you were able to secure sufficient funding, the powers-that-be would reinstate the program? If so, you may be able to persuade local and/or national potential contributors of funds, materials, products and/or services to make such donations.

For the relatively small amounts to the larger companies, and especially smaller or even struggling companies, having their company associated with a successful project undertaken by an elite university could reap significant benefits for their PR and marketing endeavors. If some of your team could put together a project website to help get project exposure established, I would think that would certainly help. Once such a website was established they could also produce a mock page or so, highlighting the company(ies, but individually) integral involvment with the project - this would help give them a "concrete" perspective for evaluating their prospective commitment to the project. Who knows, it just might work!