It is currently Wed Apr 23, 2014 4:04 pm



Reply to topic  [ 73 posts ]  Go to page 1, 2  Next
 The Simulacrum - The Altenative to an ODT 
Author Message
Certif-Eyed!

Joined: Sun Mar 25, 2012 12:33 pm
Posts: 634
The project (in its current form) is dead. :cry: It was too dangerous and unwieldy.

Small Update (8/28/12) - Just Code:
http://pastebin.com/Z2vHgcRQ
Rotary Encoder is ONLINE

-----------------------------------
Update (8/13/12) - Limping Freeman:



Lost one foot (for the day while the glue dries) because the turning was too sticky, but overall, the concept works more or less how I envisioned it.
After using it in its most basic form, I can safely say that this will be the (if not close to it) definitive consumer VR experience with Head tracking, Gun tracking, and the ability to walk backwards (w/ treadmill). :lol:

I can't wait for the Rift and its SDK! It's disconcerting to not be able to use my head to follow objects when my virtual body turns. Heck, it's disconcerting to not use my hands at all (and liberating)!

-------------------------Old Post------------------------------------------------------
Hello! I've been posting intermittently here for a couple of months, and I thought that it would be a good time to make my first contribution. Over the past few days I've started building a basic foot platform that is capable of taking a user's ankle rotations, and translating them to the virtual world via mouse emulation.

The Device:
In short: With modified rotating foot platforms attached to one's feet, one can walk on a treadmill and execute the natural ankle turning motions, without straying from the straight treadmill. These motions are then captured via potentiometer, and applied in mouse movements to the virtual world. This "sidesteps" the omni-directional treadmill problem by nipping the turning problem in the bud: the ankle.
This is the original concept I described here a while ago: http://www.mtbs3d.com/phpBB/viewtopic.php?p=75840#p75840

Pros:
  • Cheap (<$100)
  • Captures intuitive turning movement of legs.
  • Preserves natural walking motion.
  • No latency, extremely precise
  • Compatible with all mouse controlled FPSs
  • Plug-N-Play: Works just like a mouse, needs no drivers.
  • Open-Source: Everything a layman need to build it is right here, or readily available on the internet.

Cons:

The Progress:
I've been waiting to make a thread for this until I had something to show off, so here's what I have done:
An Arduino Leonardo is hooked up via RJ45 patch cords to two detachable foot platforms, each with a pressure sensor and a rotary potentiometer. It is held together with scotch tape and superglue.


The Future:
The next steps are to ruggedize every joint for the wear and tear associated with walking, and to attach a pair of shoes to the plates (More superglue in both cases? (or reusable zip-ties for shoe mounting)). Eventually, I'd like to get the pressure sensors to also detect walking (pressing W), and jumping (pressing space), especially since these require no further hardware modification.

I called it the Simulacrum because it's an actual word, and because I feel that it captures the essence of the device: 1. It's a simulation of walking. 2. It's the fulcrum between the real world and the virtual world.

It's weird that I'm making this, since I own neither an HMD nor a treadmill, but I feel like this idea has some promise and I'd hate to let it die when it has so many intuitive advantages over other solutions like the Cybercarpet, the Cyberwalk (this, by itself, is sub $100 compared to >~$100,000), and the Wizdish. When it's in a good enough state, I'm going to be taking it to my local gym and walking around in Skyrim a bit. I doubt I'll ever develop it anywhere near enough for something like a kickstarter, but I will post the plans and source code online as I build it so anyone with any amount of experience can build their own!

What do you think?


Last edited by zalo on Wed Nov 07, 2012 11:26 am, edited 6 times in total.



Sat Jul 21, 2012 12:41 am
Profile
Sharp Eyed Eagle!
User avatar

Joined: Tue Feb 21, 2012 11:57 pm
Posts: 424
Location: Irvine, CA
Do you have any ideas yet for maintaining user orientation? One other thing that I have been thinking about with regards to this idea is that it will be difficult to replicate the natural roll from heel to toe of the natural walking motion. Perhaps some kind of ball joint could give some more freedom for the user's ankle, but then you've got support issues..


Sat Jul 21, 2012 1:27 am
Profile
Certif-Eyed!

Joined: Sun Mar 25, 2012 12:33 pm
Posts: 634
I mentioned in the other thread that bungee cords through your belt loops and attached to the side railing would hopefully keep you pointed forward. From my initial tests, this is a very low friction system. I was surprised how "forward facing" I could stay without touching anything, and I think with the bungees as guidance toward the center of the treadmill more than actual forceful reorientation, it could be very comfortable and safe.

I don't have a shoe attached yet, but from what I can tell, it will feel like platform shoes. If needed, I can do what I described in the earlier thread and stick the rubber/foam shoe treads on the bottom. Hopefully that will help with traction and rolling foot movement without rolling your ankle.


Sat Jul 21, 2012 7:20 am
Profile
3D Angel Eyes (Moderator)
User avatar

Joined: Sat Apr 12, 2008 8:18 pm
Posts: 10730
Looks promising, but I still don't quite understand how you are stopping the user from physically turning around in the real world. Other that that it looks good.

_________________
check my blog - cybereality.com


Sat Jul 21, 2012 11:49 am
Profile
Certif-Eyed!

Joined: Sun Mar 25, 2012 12:33 pm
Posts: 634
Well, I put some rubber bands on them to keep them from turning too far, I affixed the shoes to the device, and I added a forward functionality so it presses w for 1/4th of a second every time you lift your foot.
Image
But I was too excited to upgrade the glue joints.

I immediately went to playing quake with it. I broke a rubberband and detached a potentiometer.
Easy to repair, but I'll need to wait to repair it until I get ahold of some stronger/more flexible glue than scotch tape and super glue.

It was kind of encouraging for the short time that I did play, though. I can easily see myself stomping through the corridors in Quake with some MUCH sturdier shoes and a treadmill that matches my speed. When it worked, it was a great way to change my direction without using my hands. Plus, the pressure sensors worked great!

For those who like seeing how poorly I code: http://pastebin.com/uXjqLd98


Sat Jul 21, 2012 1:51 pm
Profile
Petrif-Eyed
User avatar

Joined: Sat Sep 17, 2011 9:23 pm
Posts: 2167
Location: Irvine, CA
I missed the other thread, but glad I caught it this time around. It sounds interesting. The video shows you turning the camera according to the orientation of the shoes. So as you start turning on the treadmill it will rotate your direction - but then what? If you face back forward won't it turn the camera back to the original direction? How do you maintain the new course?


Sat Jul 21, 2012 3:15 pm
Profile
Certif-Eyed!

Joined: Sun Mar 25, 2012 12:33 pm
Posts: 634
The pressure pads on the shoe determine whether the foot is touching the ground. When the user lifts his/her foot, the rotational information is discarded.

It's meant to capture the movement in stride. :)

EDIT: Dangit, glued one of the potentiometers solid, have to replace it.


Sat Jul 21, 2012 4:35 pm
Profile
Sharp Eyed Eagle!
User avatar

Joined: Tue Feb 21, 2012 11:57 pm
Posts: 424
Location: Irvine, CA
You'll need to create some kind of a slip-ring, to account for the user walking in a circle or something. They are unfortunately kind of a hassle to source at this size, as I've found out from some other projects I've worked on.


Sun Jul 22, 2012 1:31 pm
Profile
Certif-Eyed!

Joined: Sun Mar 25, 2012 12:33 pm
Posts: 634
I'm trying to solve that with the pressure sensors and some rubberbands. When the user lifts up their foot, a rubberband pulls the platform back into the best position. The pressure pad is released, and the turning data from the "pull-back" is not used, so players can "ratchet" turn in a circle. It would still be a problem for a ballerina simulator where you twirl on one foot though.

In other news, I glued my other potentiometer solid in too :(
I only have one left before I have to order replacements (not expensive, but it takes a while to ship from Sparkfun).


Sun Jul 22, 2012 2:39 pm
Profile
Sharp Eyed Eagle!
User avatar

Joined: Tue Feb 21, 2012 11:57 pm
Posts: 424
Location: Irvine, CA
Ah, that is a neat workaround. It introduces some lag while the platform returns to its original position though, as well as some angular momentum and mechanical wear.

EDIT: Have you looked at Digi-Key? They get stuff to me in a day or two.


Sun Jul 22, 2012 3:42 pm
Profile
Certif-Eyed!

Joined: Sun Mar 25, 2012 12:33 pm
Posts: 634
Lag's no issue since your foot is lifting for the step while it returns.

Also, hopefully dowsing something in Acetone and letting it sit will dissolve the trouble gorilla glue that binds my last potentiometer. :(


Sun Jul 22, 2012 4:34 pm
Profile
Certif-Eyable!
User avatar

Joined: Sat Dec 22, 2007 3:38 am
Posts: 988
Cool project, but I'm wondering how you are overcoming an issue.
When you are walking and turn, you naturally cross over you legs. For example, turning left, you put your right leg to the left of where your left foot is pointing. How do you prevent this from causing you to start walking on the left of the treadmill?


Sun Jul 22, 2012 5:44 pm
Profile
Certif-Eyed!

Joined: Sun Mar 25, 2012 12:33 pm
Posts: 634
It will be weird at first, but there is a mode of walking that doesn't require strafing. When you walk and turn, sometimes you just turn your ankle/leg and it shifts your whole frame (so the rest of your body stays still) to turn in a new direction. When wearing the HMD, you won't be able to see that it is not your frame/torso turning, but just your leg.

"Then you'll see, that it is not the spoon that bends, it is only yourself."
-Creepy Spoon Kid from the Matrix

Strafing will remain an issue, unfortunately, but you can hopefully consciously override the desire to strafe with the aid of the waist bungees giving you passive tactile tension as to your current position on the treadmill.


Sun Jul 22, 2012 8:06 pm
Profile
Petrif-Eyed
User avatar

Joined: Sat Sep 17, 2011 9:23 pm
Posts: 2167
Location: Irvine, CA
zalo wrote:
Strafing will remain an issue, unfortunately, but you can hopefully consciously override the desire to strafe with the aid of the waist bungees giving you passive tactile tension as to your current position on the treadmill.


I wouldn't worry about strafing. It's an awkward movement and doesn't occur much in real locomotion. In real life, strafing is almost always the act of walking forward with your head and/or torso turned sideways.


Sun Jul 22, 2012 8:39 pm
Profile
Golden Eyed Wiseman! (or woman!)

Joined: Mon Jun 22, 2009 8:36 am
Posts: 1408
Location: Stockholm, Sweden
Cool! I would like to see a video with PiP with you using that :D

By the way, you should write a FreePIE plugin for this so you can have more generic support ;)

edit: brantlew, In a real life tactical situation strafing is used to some degree for example when firing from a fixed cover and need to move out to get a clear shot etc...

_________________
FreePIE
My blog


Mon Jul 23, 2012 4:43 am
Profile
One Eyed Hopeful

Joined: Thu Jun 21, 2012 7:23 am
Posts: 35
Would setting the treadmill on a rotating pad help? In the end, if the foot rotation is "read" as a body pivot, and that fed through to a HMD camera view pivot, then that should fool your senses into thinking you have rotated/pivoted. Yes, it takes some learning, but would boil down to "driving" your avatar with foot rotation pivot steering, and stepping for forward/backward locomotion.


Tue Jul 24, 2012 2:42 am
Profile
Petrif-Eyed
User avatar

Joined: Sat Sep 17, 2011 9:23 pm
Posts: 2167
Location: Irvine, CA
A few thoughts on treadmills.

- A simple option would be to use a manual treadmill instead of a motorized treadmill. That way you can bypass the need to control it, but there will be a lot of starting resistance and stopping inertia to deal with. There is even a type of curved manual treadmill that is easier start without rails (but expensive)




- As far as stabilizing attachments go...I ran across this video a long time ago that shows just how much stabilization and resistance you can create with some bungie cables. Nice that you can support full-on running (although doubtful you run very fast with rotating platform shoes)



- Even with cords it might be dangerous on a treadmill with an HMD - and if you fall you're likely going to smash your face right into the control panel. For added stability, safety and confidence, it might be best to just use the rails. Then you wouldn't need all the cords and you could initiate complex motions simpler and with more confidence. So why not just attach some button controllers to the rails and keep your hands on the rails at all times? Maybe even one of these big rails for added safety?

http://www.progressionfitness.net/prodimages/Treadmills/SEN.jpg


Tue Jul 24, 2012 8:00 am
Profile
Petrif-Eyed
User avatar

Joined: Sat Sep 17, 2011 9:23 pm
Posts: 2167
Location: Irvine, CA
For some reason my last pic won't post on that message???? Here it is

Image


Tue Jul 24, 2012 8:04 am
Profile
Certif-Eyed!

Joined: Sun Mar 25, 2012 12:33 pm
Posts: 634
I've looked at treadmills like the Woodway Curve, and I've decided that the cost (~$7,000) is prohibitive, and the fact that it would be no better than a virtusphere makes the reward too little. As you mentioned, the inertia and the fact that you always feel like you are walking uphill are the two evil factors here. (plus, in a virtusphere, you can strafe!).

I'd much rather use a sensory system like this: http://jap.physiology.org/content/95/2/838.full
The bungees will need less tension to detect in a system like this though.

Bars are great, so long as the system can fully operate without relying on them. When you (inevitably) trip, I want them to be there for you.

I'm just having a hard time finding a cheap treadmill that will accept computer input.

That, and my prototypes keep breaking (expanding/foaming glue should never be used near and joint of any kind).
Heh, they are either too flimsy, or too "solid".


Tue Jul 24, 2012 8:18 am
Profile
Sharp Eyed Eagle!
User avatar

Joined: Tue Feb 21, 2012 11:57 pm
Posts: 424
Location: Irvine, CA
You're not going to find a cheap one that will accept computer input. However, all of the treadmills that we make at work are controlled by sending serial data from the console to the motor controller, which generates a PWM for the motor. The easiest thing to do would be to coopt that serial connection. I will see if I can check out our code to find the format of the serial data for our products, but alternatively, if you find a good deal on a treadmill, you can reverse engineer it.

To reverse engineer computer control, you basically just need to find the serial data lines coming from the console and put them on an O-scope. Vary the speed with the console and monitor how the signal changes; that's your speed control data. Or you can find the data sheet for the processor on your motor controller and see which pin it outputs the PWM to, then spoof your own PWM with a microcontroller.

I will check for specifics at work this week; obviously I can't send you full schematics, but I will find out where to inject your PWM at least. PM coming your way with more specific brand info.


Tue Jul 24, 2012 8:46 am
Profile
Certif-Eyed!

Joined: Sun Mar 25, 2012 12:33 pm
Posts: 634
As you might have guessed, I am seriously under-equipped, so walking the PWM route is a very attractive option for a man without an oscilloscope (even though it sounds like a lot of fun).

I'm also concerned if the controller has any safety mechanisms (for the motor) or programming that will prevent the treadmill from sudden starts if I try to go the Serial way.

Would I be making a mistake in looking into a cheap one like this?
http://www.searsoutlet.com/T5-5-Treadmi ... onType=all


Tue Jul 24, 2012 9:31 am
Profile
Petrif-Eyed
User avatar

Joined: Sat Sep 17, 2011 9:23 pm
Posts: 2167
Location: Irvine, CA
Since you're investing time in treadmill systems, I thought I might throw another simple idea out there to chew on - even simpler than the shoes.

What if you just mounted a gyroscopic/compass tracker on your chest or back. Instead of using the ankle position to initiate turning you could use the twisting of your torso to turn. This would be independent of the head which could turn the camera but not the direction of motion. It's a bit of a compromise and not exactly like walking, but neither are the shoes. And it might be easier to implement.


Tue Jul 24, 2012 9:47 am
Profile
Certif-Eyed!

Joined: Sun Mar 25, 2012 12:33 pm
Posts: 634
John Carmack talked about degrees of separation in interaction, and how console players are at a disadvantage because they have to integrate velocity over time, as opposed to PC/mice players who get direct mapping between the virtual and real world.

I don't suspect that he would approve of a compromise :)

Hopefully though, the stationary treadmill approach will enable more advanced trackers like the hydra for head and hand tracking.


Tue Jul 24, 2012 10:35 am
Profile
Sharp Eyed Eagle!
User avatar

Joined: Tue Feb 21, 2012 11:57 pm
Posts: 424
Location: Irvine, CA
No, one of those NordicTracks should work fine. I'll see if I can get any info on it.


Tue Jul 24, 2012 12:42 pm
Profile
Golden Eyed Wiseman! (or woman!)
User avatar

Joined: Tue Feb 12, 2008 5:22 am
Posts: 1512
Great work zalo. But how about instead of the big clunky rotating foot discs and pots, just use gyro sensors on tops of your shoes and pressure pads on the bottom? You still get the rotation data nice and neat without all the hardware?

_________________
"I did not chip in ten grand to seed a first investment round to build value for a Facebook acquisition."
Notch on the FaceDisgrace buyout.


Mon Jul 30, 2012 7:40 am
Profile
Certif-Eyed!

Joined: Sun Mar 25, 2012 12:33 pm
Posts: 634
The true purpose of the foot discs is to actually neutralize ankle rotation without friction, the measurement method is unimportant.
If there wasn't an exposed screw that I could mount the pots to, I'd be forced to use something like a gyro.

Plus, Gyro sensors are ~45x more expensive than potentiometers (can't beat one dollar!). And I'd need one for each foot!

Also, sorry for the lack of updates recently, I ordered replacement parts and now that they have come in, I can get back to work.

EDIT: For those of you still checking this thread, good for you! Life has caught me off guard for the last few weeks, but I'm going to make an update soon, hopefully when I have it working on a treadmill.

In the mean time, here's a picture of the (painfully obvious in retrospect) velcro shoe mount with Silicone glue'd pot. Both of the joints are much more flexible now, so they hopefully won't break under wear and tear (and now I can get at them if they do!).
Image

I don't have anything else to show, except to say that walking around with this is actually a whole less awkward than I had dared to hope. The actual function of the device works much better than I could have hoped. The forward momentum of walking really helps, so a treadmill trial is (still) the only true way to tell for sure whether this is viable or not in terms of control accuracy. You do seem to be able to stay surprisingly aimed relatively forward without the need for bungees when walking too.


Last edited by zalo on Sat Oct 27, 2012 5:13 pm, edited 1 time in total.



Mon Jul 30, 2012 8:06 am
Profile
Certif-Eyed!

Joined: Sun Mar 25, 2012 12:33 pm
Posts: 634
Sorry about being incredibly slow at updating. I'll be honest, not a lot has been happening on the Simulacrum front (just got a PS3 eye and am experimenting with point tracking in processing).

I'm still too scared to bring it to a Gym for treadmill testing, but walking in place tests have been promising:



I'd like to be able to walk backwards, but it's a minor gripe considering that this was all hands-free.
The Rift will really enrichen this experience with head tracking, so you can still follow objects while turning.

Also, one of my platforms is out of commission until tomorrow (glue again! of all the things!)

In other news, a fun new game-mechanic would be limping when you get shot in one of your legs. All it would have to do is deactivate one of the feet! (It's a feature, not a bug.... right?)


Mon Aug 13, 2012 7:08 pm
Profile
Petrif-Eyed
User avatar

Joined: Sat Sep 17, 2011 9:23 pm
Posts: 2167
Location: Irvine, CA
Interesting. You've made some good progress. Two thing that occur to me.

1) It seems like it would feel more realistic if you just set the game to walking speed and held a continuous walk instead of starting and stopping on each stride.

2) I'm still not sure how the turning mechanism would work in stride. Right now it sort of looks a bit like a foot joystick. Does it feel intuitive (like walking) or does it feel like a control?


Mon Aug 13, 2012 7:24 pm
Profile
Sharp Eyed Eagle!
User avatar

Joined: Tue Feb 21, 2012 11:57 pm
Posts: 424
Location: Irvine, CA
Good work, Zalo. I can see this approach having potential once you smooth out the edges. Have you done any (short-distance) walk-around tests?


Mon Aug 13, 2012 7:46 pm
Profile
Certif-Eyed!

Joined: Sun Mar 25, 2012 12:33 pm
Posts: 634
Thanks! I have walked around a bit with both shoes (nothing connected) and I can vouch that it can feel natural. (It feels good enough to be immersive when I close my eyes (and its a terrible idea to walk around in the world on these with your eyes closed!))

Also, I guess I wasn't too clear on this before (sorry): because there was only one foot, it only detected the movements of one foot. I made it so it holds down the w key for 1/4th of a second whenever you lift a foot because that's usually how long a stride is. When I get it in a more functional state with two feet, it should come to resemble continuity in stride.

One thing that is kind of annoying right now is that it always assumes you want to move forward whenever you lift up your foot. This will be a non-issue on a treadmill with positional sensing, but I'll have to think of a compromise for the near future. I may be able to achieve something like this when I get my PS3 optical tracking up to snuff.

My aim is to get it to feel as natural/immersive as any of the other components that make up a VR apparatus, and I think I can do it with this design.


Mon Aug 13, 2012 8:09 pm
Profile
3D Angel Eyes (Moderator)
User avatar

Joined: Sat Apr 12, 2008 8:18 pm
Posts: 10730
Cool man! Looking good. Still seems like turning is a little awkward, but maybe you can smooth that out in the future.

_________________
check my blog - cybereality.com


Mon Aug 13, 2012 8:48 pm
Profile
Golden Eyed Wiseman! (or woman!)
User avatar

Joined: Tue Feb 12, 2008 5:22 am
Posts: 1512
I really like this zalo good work. The pressure switches activating the 1/4 walk is a great idea. You could even time their interval to activate a run key, if not timed to analogue walk speed as some engines support analogue joypads.

_________________
"I did not chip in ten grand to seed a first investment round to build value for a Facebook acquisition."
Notch on the FaceDisgrace buyout.


Tue Aug 14, 2012 4:20 am
Profile
Sharp Eyed Eagle!
User avatar

Joined: Tue Feb 21, 2012 11:57 pm
Posts: 424
Location: Irvine, CA
The only intrinsic difficulty I see with this idea is that you can't "lean" into your turns, the way you would be able to in a free-walking scenario. But that will be a fact of life for any stationary omni-motion solution. In this case though, your feet are also on rotating pads, so I wonder if it will cause users to lose their balance until they have spent some time learning how to use this system.


Tue Aug 14, 2012 1:02 pm
Profile
Certif-Eyed!

Joined: Sun Mar 25, 2012 12:33 pm
Posts: 634
Just tried it with both feet, and I have to say, it's substantially harder to control for one reason:
Quote:
The pressure switches activating the 1/4 walk is a terrible idea.


The turning itself was no problem. It's turning, getting off by a few degrees, and having to lift your foot to do the correction-turn, and accidentally moving forward off of a cliff. I shouldn't have to strategize my next step this much if I want to be immersed!

I've taken the concept as far as I can without any extra hardware, but I'm going to need to decouple the foot lifting from turning if I want to be able to achieve finer control (such as turning in place :p)

I've also decided that a manual treadmill will actually be as good as the automatic one in terms of resistance because:
1. On the automatic one, you walk against the supporting bungees to move forward on the treadmill so it can move you back.
2. On the manual one, you walk against the supporting bungees to move the tread back.

There will always be a constant pressure against walking with either one, so I figure that at least the manual one is significantly cheaper, and I can measure speed and distance accurately using a mouse on the tread.

Unfortunately, what this means is that I'm putting this project to sleep again until I buy a manual treadmill and a few other things. Sorry!

EDIT: Treadmill and rotary encoders are en route.


Tue Aug 14, 2012 1:27 pm
Profile
Sharp Eyed Eagle!
User avatar

Joined: Tue Feb 21, 2012 11:57 pm
Posts: 424
Location: Irvine, CA
What do you mean by 'rotary encoders'? Something more reliable than pots?

EDIT: Nevermind, Googled my own question.


Tue Aug 14, 2012 3:16 pm
Profile
Certif-Eyed!

Joined: Sun Mar 25, 2012 12:33 pm
Posts: 634
EDIT: Oops, accidental post.

Uh, the Digispark kickstarter is looking really cool. I think the next iteration of this project will use that as the board. It will brings down the total cost of this whole apparatus by $20 (a non-trivial amount, considering that the whole thing was ~$50 to begin with (BYOTreadmill))!

Also, I guess I should clarify, the rotary encoders' main feature is that they can rotate full 360 (as opposed to pots which are like 270 or so), but the cost is that they only have 12 detectable positions. Since they'll be used to measure treadmill distance and speed, high resolution isn't super important (and they're still fairly cheap at 3 or 4 bucks a pop). Using these, as opposed to a mouse on the tread surface, allows me to unify the whole system into one arduino, and to keep an accurate record of total distance traveled over time (should I calibrate the softwareto include that functionality).

Anyway, I'm still optimistic! I saw the wizdish got a shout out by Palmer on his kickstarter, and I know my system is better with just the ability to walk backwards, turn in place, and the walking motion itself. I can't compare the overall cost yet (does anyone know how much he's selling it for?), but I think my design is competitive on that front too.

EDIT: Hah, I just realized, if I sacrifice just a little realism, I can add the ability to strafe by stepping off the side of the treadmill without even modifying any of the other features. Will be kind of (extremely) difficult to strafe while walking, but it will make the ability to strafe around corners much easier. What matters is being able to move using just your feet intuitively so your hands and head can stay free for the real immersive stuff.

EDIT: I can imagine someone who is tired of walking forwards, so they turn their head to the right, and step off to the right to strafe in that direction as an alternative to walking the hard way. Would probably be useful in skyrim, lol.

EDIT: I've come into a large quantity of foam tiles, which, with clever carving and gluing, would make each step feel more like a roll instead of a "clop" like a snowshoe. It would act as the curved underside of the shoe (the "rubber" part of the sole)?


Thu Aug 16, 2012 12:17 pm
Profile
Certif-Eyed!

Joined: Sun Mar 25, 2012 12:33 pm
Posts: 634
The Treadmill and have Encoder arrived. The treadmill is a little sticky, but bearable (mostly static friction).
It also had no place where encoder could be attached, so I had to get a little creative:
http://vid34.photobucket.com/albums/d144/Zalo10/Omni%20Directional/IMG_0591.mp4

The whole setup looks alright, though:
Image
The platform that the laptop is sitting on is just for looks and debugging.
The incline will be a thornier issue (fantastic for walking forward, terrible for backward).

Ran out of solder, need to buy some more to wire it up and get data out of it.

EDIT: Remind me to never EVER get lead-free solder again. That sh*t is terrible!

Also, my brother just broke my quarter setup. So much for creativity!


Last edited by zalo on Sat Oct 27, 2012 5:19 pm, edited 3 times in total.



Fri Aug 24, 2012 3:31 pm
Profile
3D Angel Eyes (Moderator)
User avatar

Joined: Sat Apr 12, 2008 8:18 pm
Posts: 10730
Looking good man. Been wanting to get a treadmill for a while now, I imagine it's fun.

_________________
check my blog - cybereality.com


Fri Aug 24, 2012 3:57 pm
Profile
Certif-Eyed!

Joined: Sun Mar 25, 2012 12:33 pm
Posts: 634
No pictures this time around, the setup looks almost precisely alike the previous picture (except for the repaired Rotary Encoder and some spacers (books ;) ) under the near end to make it level).

The whole treadmill measuring portion of the system is up and running, and working quite nicely. The Rotary encoder (despite having only ~11 measuring positions per turn) is extremely sensitive. It works both forwards and backwards as well, which is a nice plus. For those who like math and calibrations, the Rotary Encoder gives me 1301 ticks (~11 ticks per turn) for a full 41" turn of the belt. The circumference of a quarter (the gear I'm using lol) is exactly 3". So it doesn't match up exactly (due to small alignment errors of the quarter to the axle of the encoder) but its VERY reliable (second spin at high speed gave me exactly 2605; even that small error is most likely sloppiness on my part).

The current code simplifies the speed of the treadmill into 3 speeds: Slow (w+Alt), Normal (w), and Fast (w+Shift). The code was comparably easy to make, however there is a 100ms latency while the sensor accumulates turns to calculate speed. It is unnoticeable though and can be shortened if it becomes an issue. Speaking of current code, in the interest of Open Source, here it is: http://pastebin.com/Z2vHgcRQ This code only works forwards, but adding in reverse functionality would be simple for anyone who knows Java (after reading the code). I'll add it in as the project matures.

I am now heavily leaning (hah!) toward the idea of using just one foot measuring platform at a time. It's easier to stand and easier to turn. By virtue of how I built it, I can switch to a dual system at any time (easy as plugging in an ethernet cable), but I think I will stick with the mono version for some time to come.

Also, mad respect to all of you hardware builders out there. Building hardware is great until you find you're missing that ONE part that you need, and mistakes are expensive and extremely time consuming. No Ctrl-Z and no downloading what you don't have. Just Patience and Tenacity. You guys are awesome. (Same, I imagine, as doing an IAMA on reddit! Thanks for that, Palmer!)


Last edited by zalo on Tue Aug 28, 2012 9:21 pm, edited 1 time in total.



Tue Aug 28, 2012 8:35 pm
Profile
Petrif-Eyed
User avatar

Joined: Sat Sep 17, 2011 9:23 pm
Posts: 2167
Location: Irvine, CA
I'm glad you're still working on this. Continued development usually means that the project shows promise.

zalo wrote:
however there is a 100ms latency while the sensor accumulates turns to do calculate speed. It is unnoticeable though and can be shortened if it becomes an issue


Motion system have plenty of problems to deal with, but one thing we're lucky about is latency tolerance. From mental decision, to weight shift, to stepping is a long and involved chain of events and it's hard to pin down exactly where the walk begins. I find that even 1/4 second delay still feels responsive for normal walking.


Can you give me a tip? What are you using again for a foot-down indicator?
Edit: Ah, I see it now - that Sparkfun FSR. Can you then give me a quick run-down of your FSR integration method?


Tue Aug 28, 2012 9:08 pm
Profile
Display posts from previous:  Sort by  
Reply to topic   [ 73 posts ]  Go to page 1, 2  Next

Who is online

Users browsing this forum: No registered users and 2 guests


You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot post attachments in this forum

Jump to:  
Powered by phpBB® Forum Software © phpBB Group
Designed by STSoftware.