Introducing the STALKER! (project ended)

Let's Make Robots Projects Showcase
21

To be quite honest im stabbing the dark with this one.
i know that styrofoam can be vacuum formed and can
withstand temp’s up to 75°c.
the idea is to simply lay the styro part on the vacuum
bed face up. i only intend (at this point) to form a
shell to cover the face of the leg.

once the plastic has cooled i will trim off the excess,
leaving (hopefully) an outer casing, of which will
be hollow. like an RC race car shell!!

22

Oh, this sounds really cool, where can i get myself one of these vacuum molders? (just like a small one though)

23

i dont acually own one but i am looking into
buying one if i take it further!
i plan on sending the part off to be formed.
also i have a good friend that may know of
someone who has one.

yeah would be cool to have one tho!

24

link to vacuum formers
img187.imageshack.us/img187/8373/vacuumformingprocesspg6wl9.th.gif

castcraft.com/cat_vacplans.htm

25

It will be cool to see the bot get armor!

Good luck on the vacuum forming!

26

Well, i was lookin for a vacuum former i could buy, not buy the instructions for. Good link though. Ill do a homemade one but the ones in that link looked like it needed to much.

27

What are the demensions of the foam shapes? Are they meant to fit the SES? They look really big compared to your floor tiles. :confused:

28

kitchen work surface!!

these shapes have only been made by eye to hand. so its approx the right size. bearing in mind im not adding the styrofoam to the leg and once the shells have been moulded they will be hollow so the servos etc will be hidden insde!

this image wouldnt load so i have a link to it!

you can see the size difference and placement of shell, i still have some mods to do to them still so diamentions will chance slightly!

[img=http://img301.imageshack.us/img301/2929/20080323101943el5.th.jpg]

29

ahh, thought those were tiles.

30

i have found a video of the stalker in action!
here you can see the gait involved!

uk.gamespot.com/video/928399/616 … me-guide-6

31

Well the video clip makes the gait look easy!

Simply lift and move two legs at a time, while the other two advance the body. And I notice that you have 3DOF legs, apparently matching those in the video.

That said, I must observe that there is zero stability while up on those two legs! This is basically like a biped, two pairs of legs moving together. Can it be done? That might be a trick. If one believes the gait in the video (animation), then simply reducing the legs of the CH3-R BASIC Atom code from 6 to 4 would do it! However I’m not ready to accept that.

You could modify the code, and suspend the 'bot from the servo horn that the head servo is mounted on. Then watch the legs!

OR, you could implement a wave gait as in the video you (I think) posted of the MiniQuad II 'bot (might have been in another thread). This is the 4-state gait mentioned. Also called a “creep” on the Oricomtech.com website.

Alan

32

I noticed the same - the whole thing balances on two single-point feet while the other two lift, reposition, and land. Easy to do in a game, not so much in reality. I won’t say that it’s impossible, but I’d put it firmly into the “difficult as hell” category, to keep the body stable, and do it without dragging at least one foot in the process.

On the other hand, there seemed to be a few moments in the clip where it utilized a ripple-like gait in what looked to be a figure-8 pattern, though that wasn’t the dominant locomotion displayed.

Looking good, though. I’m interested to see how the vacu-forming works out, and how it works out when all is said and done. Looking great so far!

33

Couldn’t the balancing be controlled by shifting the wieght far above the stationairy feet and than have the weight fall towards the feet landing forward? Something like the tripod gate being presently being done?

34

The tripod gait is dynamically stable, it doesn’t need to shift weight. Bipeds need to shift weight. Humans “fall forward” and “ketch” themselves. That’s fairly tricky for a 'bot to do.

I did just come across a biped 'bot that used acceleration sensors to keep balanced, but the legs weren’t advanced enough to get the 'bot walking.

randy.gamages.com/vito.shtml

As I mentioned the animated 'bot in the video uses legs like a biped (legs paired up), so probably biped walking “tricks” would apply.

If the legs were more agile (more joints?), then possibly they could “kick off” and fall forward, similar to a biped. But that might limit it to travel exactly perpendicular to the two stationary legs; otherwise the 'bot would kinda fall a little sideways, rather then in the line of desired travel. But this is all speculation, as I haven’t programmed a quad.

Well, you could always move like this:

pdjinc.com/video1.m1v
pdjinc.com/video.m1v
pdjinc.com/video3.m1v

From the Arachnobot website:

pdjinc.com/arachno.html

I wasn’t impressed.

Here’s a paper on a quadrupod robot kit assembly. there is a good explaination of the quad gait near the end

robotshop.ca/PDF/rbmci04_manual.pdf

Alan KM6VV

35
36

due to the lack of communication (replys) from a vaccum forming company i have abandond sending them emails and i had decided to fabricate the leg shells out of my RC Helicopter i picked up from a market last week.

img528.imageshack.us/img528/1687/112106628obp8.th.jpg
img91.imageshack.us/img91/3601/cabqw1.th.jpg

once i took the body cabin off i simply cut the carbon fiber case with a stanley blade, (worked very well to).
then i had some black spray paint laying about.

once dried i has some of the nylon stand-offs that i’v used to pull the sides closer so its not so intrusive.
also this has helped to make it stronger!
img528.imageshack.us/img528/8885/20080516201818id0.th.jpg

img234.imageshack.us/img234/7204/20080516202601ac6.th.jpg
img524.imageshack.us/img524/7267/20080516203059vw4.th.jpg
it then slotted into position very well.
obviously it doesn’t look as first intended but i was bored of waiting!
i have another 3 body cases on order so expect more…!
any ideas on improvements?

37

Hi,

might i add to the quadropod gait discussion?

You could solve it as i did in my first approach on a hexapod gait, where i first moved three of the legs (two in your case of course), and didn’t move the body itself until all legs were on ground again. Then when the body has moved the length of the step, the next two legs moves.

Here’s a video of the gait in action (the best view of the gait is on the laptop):
video.google.com/videoplay?docid … 2QLH943gCQ

This may not solve the problems completely, but may help you to it.

Anyway, your projects looks wicked awesome! I was thinking about making a nice shell of styrofoam, but didn’t get to it, so it’s nice to see someone else do it (though I guess they will only work as a template for the vacuum forming right?).

38

the complexity of a quadruped far out weigh that of a hexapod. i have worked on a few hexapods including the EH3-r AH3-r and a few iv designed in the past.

With a hexapod tripod gait, it’s possible to make successive triangles overlap so that the CoG is always within a triangle.
That makes walking much more tolerant of mistakes.

With a quadruped tripod gait, it need a combination of successive triangles. So the CoG has to leave one triangle and enter the
next one just as the old triangle is disappearing and the new triangle is
forming. That’s tricky to get right every time.

If the CoG is a little too far forward then it leaves the old triangle
before the new triangle has formed. The robot falls forward but,
fortunately, the new foot is about the go down and the robot falls forward
into the new triangle.

But if the CoG is a little too far backward then the old triangle
disappears while the CoG is still inside it. The robot falls backward just
as the old triangle is disappearing. So the robot tumbles over backwards.

That’s what makes the quadruped tripod gait much harder to get right than
the hexapod tripod gait.

Yeah, sounds tricky, but to add on top of that changing directions, and
it’s even more of a challenge!

39

Man, your legs on the quadrapeds are beautiful… Using that RC copter’s body repainted makes it look real wicked…

The quadrapeds gaits are pretty hard. I see what you are saying about the triangles disappearing and overlapping it… It gets more complicated as your increase the speed of the “walk”

40

i am happy you like the idea of using the body shell, and that it looks

. This idea as i say is only a concept cos fabrication of the actual leg design is becoming a pain.

only time will tell what this robot will look like.
still waiting for other body shells to arrive. (ebay)!

i am still yet to redesign the main chassis as i need to expand the distance of the servos slightly as i would like to allow for more movement in the Hip!
planning to reverse the “Off-set bracket” on the HipV and connect it direcly onto the given face of the “C bracket” of the HipH.

this will also increase the leg / leg distance, but i dont think it will look as cool!!!

thank you.