Author Topic: a very new type of robotic problem.... (Read 9356 times)

suren · « **on:** November 07, 2006, 08:29:00 AM »

hello friends,while going through one of the competitions i faced this problem..there is a wall 15(height)*100(length)*5(width)..a robot whose maximum dimensions are given (10(height)*30(length)*15(width))[ALL DIMENSIONS ARE IN CENTIMETERS]
the robot has to cross this wall by any mechanism..here all the wall climbing mechnaisms like pneumatics or anything like crawling on the obstacle fail because of the short height of the wall 15 cm..So the robot should use some mechanism like toppling over the wall maybe that is why they permitted such large length double of the height of the wall...plz suggests any mechnaisms to work upon..

Charlie1138 · « **Reply #1 on:** November 07, 2006, 08:47:57 AM »

If you only have to cross the wall once you could piossibly use a really large spring.

JesseWelling · « **Reply #2 on:** November 07, 2006, 07:28:18 PM »

I'm imagineing a robot that uses tank treads for locomotion.
But there are two sections of the tank with a hinge in the middle.
So the tank normaly trives around folded up and then deploys it's other section when it needs it.
and then it just sets the top section down on the ledge and snakes over it in a catiplliar type of fassion....
If that sounds interesting I guess I could draw up a quick schematic but remember I draw like a 4 year old.

suren · « **Reply #3 on:** November 09, 2006, 10:07:18 AM »

jesse i was looking for such kind of an answer...can you elaborate on this and a bit more stress on how do u achieve that folding structure with a robot..
regards
suren

JesseWelling · « **Reply #4 on:** November 09, 2006, 02:34:27 PM »

probably the easiest thing would be to use 2 servos as such
__________ __________
|_________}=o==o={_________|
\________/ \________/

So bare with the ascii art but 'o' is a servo and '{' is an attachment point to the chassis.
the tanks treads are on the bottom and the electonics or on the second layer so most of the
time it rides around like this:
________
/________\
|_________}=o
_L_______ [
|_________}=o
\________/

Se that L? I would make some kind of resting bracket so you
could turn off the servos and only turn them on and waste power
when you needed to.

makeing sense yet? you'd have to work with you design to make it fit but it
would be pretty mobile untill you extended the tail out.

The key to this design would be to make it realativly light (as with any design)
on both ends.

Maybe admin can coment on the mechanical feasablility of my silly design.
But if you make it you should give it a good name. Like "Cobra-Comander" or something

Admin · « **Reply #5 on:** November 09, 2006, 03:33:42 PM »

When do you need this by?

Just your luck, I was in an ASME robotics competition that does exactly what you describe. Made it to nationals too, meaning I have a few dozen pics and videos of about 20+ different robots achieving this task. Each robot had a neat way of doing it too, and for a wall twice as high as you describe.

Im really busy n all for the next few days, but Ill try and get something up on SoR that you can use . . . Ill probably compile all the videos into one and you-tube it for you.

Quote

Maybe admin can coment on the mechanical feasablility of my silly design

Based on my experience, that actually looks like a good idea. Make sure both halves can bend fully both ways, or it will get high centered up on the wall.

JesseWelling · « **Reply #6 on:** November 09, 2006, 06:44:42 PM »

awww gee-whiz....I'm blushing

suren · « **Reply #7 on:** November 10, 2006, 03:40:28 AM »

admin that will be really great..
actually i am in the organising team of a all india robotics society which is organising second largest robotics competition in india around 6 months from now on..we are thinking out on the problems..we were just drawing plans ..i saw this problem while i was participating in a robotics competition..i felt i could possibly modify this plan or add on to it something and this would be a very neat and inspiring problem..
regards
suren

JesseWelling · « **Reply #8 on:** November 10, 2006, 03:52:23 AM »

Saw this and it looked relevant even though it's not exactly like what I was thinking

http://www.superdroidrobots.com/shop/item.asp?itemid=662

suren · « **Reply #9 on:** November 10, 2006, 03:55:13 AM »

jessse i understood your folding mechanism..u mean approaching that wall and then we have that upper portion resting on the obstacle face..
then u have to go over the obstacle and u have to even take the other tank part..
then i think u even need tyres on the upper parts upper surface..(the electronics part u mentioned)then it had to pull the other part..which is heavy..
how do u achieve the motion after resting the electronics part on the obstacle..plz correct me if i am wrong
regards
suren

JesseWelling · « **Reply #10 on:** November 10, 2006, 04:22:28 AM »

um-mm maybe you missunder stood my drawing some what....we'll describe it by the layers
__________ __________
|_________}=o==o={_________| <<Electronics like sensors and the brains and of course the servos
\________/ \________/ <<Tank Treads, Motors, and batteries

So really what you are making is two tanks connected by a 2dof servo arm.
But you are just making them so they are compact enough to stack (with possible
Brackets to rest on when all folded up).

This could be done with wheels as well but you would want to use big ones that almost touch
so you don't get high centering. so you would have two bots that look like this:
http://www.budgetrobotics.com/shop/?shop=1&cart=352887&cat=156&
(but that one might not fit you dimensions)

admin, I could use your ms-paint skills....

suren · « **Reply #11 on:** November 10, 2006, 09:56:17 AM »

well jerry i grasped ur methodology..but i think it will be a bit more prbolem in balancing especially while loading the one part over the other..but i appriciate the concern shown by you in solving this problem..
regards
suren

Charlie1138 · « **Reply #12 on:** November 10, 2006, 03:33:32 PM »

Suren, I received your PM.

I don't really have it all worked out, but I would just try to mimic a flea's action when leaping. Maybe you could find some info on them on the internet -- how they work precisely. I would make sure the spring spreads the load out wide with 2 prong-like legs. And the vehicle/robot would need to be able to right itself to keep moving if flipped over.

Also.. you would need some type of servo release to activate the spring.. somekind of trigger. The first thing that came to mind was some release mechanisim they use on that TV show MYTHBUSTERS. They use some little release mechanism where you pull a string.. and it holds large weights. Maybe you could track that down. They have a message board on the Discovery channel website where they read and answer some people's questions.. you might get lucky.

I was thinking a leaf spring from a car would be good to use... maybe smaller though.. maybe from some type of wagon for a golf cart or something like that. This would be better than a coil spring, maybe? But you would have to experiment with it. I don't have any personal experiences with it, though. I'm just throwing out ideas.

Hope I helped, good luck to you.

suren · « **Reply #13 on:** November 11, 2006, 11:24:55 AM »

charlie what u suggested is of course a very good mechanism..i thought in such lines ..the only problem that i face is the releasing mechanism..
the vehicle overturns still it will run coz i can make the wheels greater in size then the rest of the body and it will work even after overturning..i did made caculations based on spring force and weight of the robot and the spring needed..only problem is how to release it..if u can find something good for this..plz do tell us..

Admin · « **Reply #14 on:** November 11, 2006, 01:50:22 PM »

so wait, the robot is only required to climb a small wall? sounds too easy . . .

the competition i did it had to climb the wall (often twice the robots height), retrieve a heavy object, then climb back still carrying the object . . . but then again we were allowed to do remote control . . .

ill get a video done by tonight i think . . .

Admin · « **Reply #15 on:** November 11, 2006, 03:26:44 PM »

Ok I went through all my videos of the other robots and they were of too poor quality to post here . . . The lighting was really bad

But I also took a lot of pics of the bots as they went over the wall, and have posted them here:
http://www.societyofrobots.com/images/ASMEcomp/

Some robots had split bodies, some laid down ramps, one had these two robot arms that lifted the bot over the wall, two had these elevator legs.

Ill post a video of the robot my team made when I can, as the quality came out better.

Charlie1138 · « **Reply #16 on:** November 11, 2006, 09:11:42 PM »

Suren, I found this release mechanism...

http://www.kk.org/cooltools/archives/001035.php

This thing is way too expensive and overkill because it is made for really high tension jobs.. but maybe we can figue out how this thing is made and reverse engineer it?

suren · « **Reply #17 on:** November 14, 2006, 10:18:05 AM »

thanks admin..i images are really good..plz post that video..videos are more helpful than the images..as u get to see the whole action
regards
suren

Militoy · « **Reply #18 on:** November 14, 2006, 02:20:08 PM »

I'm new to this site (though not to robotics) - Maybe I don't understand all of the rules of this particular challenge, but it seems to me that a simple, non-articulated tracked bot 30 cm long should easily be able to scale a flat-topped wall only 15 cm high; Especially if it happens to be built with the center-of-gravity biased forward, to help it "tip" over the wall.

JesseWelling · « **Reply #19 on:** November 14, 2006, 03:04:29 PM »

Effective but what kind of fun is that?!

Admin · « **Reply #20 on:** November 14, 2006, 03:33:40 PM »

size does matter.

Militoy · « **Reply #21 on:** November 14, 2006, 09:14:10 PM »

Quote from: JesseWelling on November 14, 2006, 03:04:29 PM

Effective but what kind of fun is that?!

I guess you're right - but I have fun working with robots, no matter how each problem is solved. Since my current challenge is DARPA's "Urban Challenge", and the robot we're working with costs cubic $$$$, I tend to concentrate on meeting each technical challenge by the most straightforward route possible - and then I move on to the next task.

JesseWelling · « **Reply #22 on:** November 14, 2006, 10:17:55 PM »

That's Reasonable. When you are making the decisions on implementation.
But if someone is just asking for ideas, I don't see any reason to stifle creativity.
It's up to them to decide what is feasible or not.

After all, if Alan Turing had never created his off the wall test,
would researchers have strivied for it and learned something in the process?

Charlie1138 · « **Reply #23 on:** November 15, 2006, 07:36:46 AM »

Suren, at the site for the Seacatch you can see how they work.. http://www.seacatch.com there is a little animated GIF showing it.

It basically works like a pair of vice grips. You could posssibly modify some vice grips to do the same thing.

Another idea would be to make a mechanism like a bolt lock on a door with a motor driven worrm gear to slide the bolt to release the spring.

JesseWelling · « **Reply #24 on:** November 15, 2006, 10:17:05 PM »

http://www.youtube.com/watch?v=weYSePU-W4A&mode=related&search=

I like this one!

Charlie1138 · « **Reply #25 on:** November 15, 2006, 10:29:50 PM »

If you wanted to go really high tech you could do it like the Ibot stair climbing robotic wheelchair.

suren · « **Reply #26 on:** November 16, 2006, 08:19:46 AM »

charlie donot u think that would increase my budget considerably..charlie i will try that ur realease mechanism..in that robt body should go in a projectile motion..so will have to realease a combination of two or more srpings..
and the tank folding one is good too..will try that one too..
regards
suren

Charlie1138 · « **Reply #27 on:** November 16, 2006, 06:48:33 PM »

Suren, haha.. yes.. those huge industrial seacatches would definitely hurt your robot budget.

I would make sure that the the 2 prongs to launch your bot are connected together.. so they only need one release. Otherwise you will have problems making them release at the same time and the bot will be hard to control. It will go all different directions.. which could be fun come to think of it.

suren · « **Reply #28 on:** November 17, 2006, 11:08:08 AM »

ya charlie..that again is a task left to be checked out..

Admin · « **Reply #29 on:** November 19, 2006, 01:24:26 AM »

ok i made a quick video and an even quicker tutorial/documentation of the wall climbing robot i made 2 years ago . . .

http://www.societyofrobots.com/robot_asme.shtml

its kinda bad documentation, but ill fix that up and make it more useful when i get more time . . . if people have questions on anything particular, ill go into more detail . . .

i also have a solid works CAD model of it (7mb) if anyone is interested.

(message modified to correct for url change)

News:

Author Topic: a very new type of robotic problem.... (Read 9356 times)

Related Topics