Subject:
|
Re: (Un)steerable Shrimp
|
Newsgroups:
|
lugnet.robotics
|
Date:
|
Tue, 2 Oct 2001 03:26:15 GMT
|
Original-From:
|
Brian Alano <ALANO@KIVA.NETsaynotospam>
|
Viewed:
|
818 times
|
| |
|
|
Elijah Meeker wrote:
> ...it seemed to me that wheel
> height was a defining aspect of the shrimp design.
It seems so from the literature, but I wonder if rocker length also
plays a part. There probably is a rocker arm length to wheel diameter
ratio beyond which increasing the rocker length doesn't do much good.
The seminal idea was to climb the obstacle one axle at a time. To me
this implies that one set of wheels should clear the obstacle before the
next set engages the obstacle. However, the SHRIMP at
http://www.bluebotics.com/products/shrimp/
seems to not require that so rigidly.
> Steering issues:
>
> [snip] I don't have any micromotors
> but I am hoping they will have suitable power for turning the turntable
> via a worm gear. If anyone is sure they won't I would greatly appreciate
> a heads up.
I'm pretty sure they won't work. They are s-l-o-w and w-e-a-k. Suitable
only for moving light loads slowly. I've used a micromotor, worm gear
and 24-tooth gear to work a pneumatic valve
(http://www.kiva.net/~alano/brian/pa/index.html). It takes 10-14 seconds
to fully cycle (5-7 seconds each way).
I recently saw a chart comparing all the Lego motors by torque and
speed. Sorry I don't have the URL, but I know it's out there somewhere
on the 'Net.
> Going to separate motors for front and back wheel raises the issue of
> sensors for each. I currently have a rotational sensor, but am thinking
> that I will likely wind up using a light sensor with an encoder strip.
> Again, any suggestions would be welcome.
Being less ambitious, I plan to use touch sensors to signal the center
point of the front and rear steering positions, and not worry about
precisely matching angles between the front and rear beyond that. (To
conserve inputs, I also plan to multiplex the touch sensors on a single
input.) Mario Ferrari managed to mechanically link his front and rear
steering, but his pictures seem to have disappeared from his web site
(after apparently taking quite a ribbing from his Italian
contemporaries, I might add.)
> Bogey Issues:
> [snip]
> The main issue with bogeys is lateral strength and stability. In this
> design I tried to address both by putting the rocker arms flush against
> the shrimp body and trying to eliminate places the arms would catch.
> This worked marginally well, but I have since headed down the path of
> putting a a single pulley as a pivot bushing and using 1/2 width lift
> arms as runners farther out.
I'm still experimenting with this issue, too. I don't have much success
to report yet.
BTW, I love the diagonals on your rear wheel support.
> Then, as that wouldn't
> work for the top I found that shoving a bit of flex system hose in a
> regular black pin (some pins allow this, others don't) I could make a
> very strong, and flush, pivot point.
Good idea
> I think that in theory there is an absolute height that a shrimp can't
> climb, that being above 1.5 times the wheel height, however it can
> usually do a bit better.
I seem to remember it being 2.0 times. Could this have something to do
with the shorter relative wheel base of the SHRIMP you modeled versus
the original SHRIMP design (i.e. smaller rocker arm to wheel diameter
ratio)?
> Still, I headed down the path of putting
> sensors on each bogey (which will be followed by one on the front wheel,
> for sensing when a given wheel assembly will likely fail to climb it's
> bit of the obstacle. However, as I need three of these at a minimum (and
> would like to know whether the problem is on the right, left or center)
> I am going to run out of sensor inputs given the need for them with
> steering. Soooo, there will be more purchasing opportunities before I
> get to a shrimp that navigates an unknown environment.
I'm very interested in what you will use as an indication of pending
climbing failure.
Again being less ambitions, I was planning to mount my Techno Stuff
DIRPD (Dual-InfraRed Proximity Detector) high in front to detect and
avoid big, tall obstacles. Well, you say, what about a rear or side
bumper? What happens when it gets stuck climbing halfway up something? I
say, I'll answer those quesetions when I get another RCX or a fast input
multiplexer (and more sensors!). I might get ambitious enough to mount
the DIRPD on an arm with a couple degrees of freedom so it can peer
around different directions. Then again, I don't have any free outputs
(Mantas, I need Mantas!), so perhaps the best I can do is attach the
DIRPD to the front steering assembly.
Thanks for the pics. The SHIMP, for all it's promise, is quite a
challenge to build. Every picture helps!
Maybe I'll have something worth posting in a month or less.
-Brian
|
|
Message has 2 Replies:
 | | Re: (Un)steerable Shrimp
|
| (...) I agree. In my SHRIMP II version I reduced the length of the front arm from 16 to 12 studs, and this affected negatively the maximum height of the obstacle that the platform can overcome. (...) Actually the rule is not so rigid. A SHRIMP can (...) (23 years ago, 2-Oct-01, to lugnet.robotics)
| | | Re: (Un)steerable Shrimp
|
| (...) I hit upon near total success last night, and the solution turns out to be very simple (surprise!). It is an ordeal for me to get pics on the web at the moment (I'm inbetween ISPs), but here it is in essence. The side of the shrimp is 4U high (...) (23 years ago, 2-Oct-01, to lugnet.robotics)
|
Message is in Reply To:
| | (Un)steerable Shrimp
|
| Hi all, I have finally gotten my steerable shrimp to it's next blocking point. The "Un" part of my subject comes about as I am still lacking a Barnes Motor Amplifier to handle the 5 minimum motor controls and two micro motors for the steering (...) (23 years ago, 1-Oct-01, to lugnet.robotics)
|
7 Messages in This Thread:
  
- Entire Thread on One Page:
- Nested:
All | Brief | Compact | Dots
Linear:
All | Brief | Compact
This Message and its Replies on One Page:
- Nested:
All | Brief | Compact | Dots
Linear:
All | Brief | Compact
|
|
|
|
