POSITTRON: yet another proposal for positional head-tracking

This is for discussion and development of non-commercial open source VR/AR projects (e.g. Kickstarter applicable, etc). Contact MTBS admins at customerservice@mtbs3d.com if you are unsure if your efforts qualify.
User avatar
drifter
Binocular Vision CONFIRMED!
Posts: 242
Joined: Tue Nov 27, 2012 9:39 am
Location: Little Britain

Re: POSITTRON: yet another proposal for positional head-trac

Post by drifter »

The nice way would be to put a few leds on the back directly on the 2 straps.
Now the tricky part would be to get their positions relative to the front leds...
PatimPatam wrote:how much the tracking algorithm is going to tax the CPU/GPU (or will it be done by HW?).
I guess it will be cheap, as "real-time" is not needed
https://twitter.com/ID_AA_Carmack/statu ... 2353905665
User avatar
PatimPatam
Binocular Vision CONFIRMED!
Posts: 214
Joined: Thu Jun 28, 2012 1:31 pm
Location: Barcelona

Re: POSITTRON: yet another proposal for positional head-trac

Post by PatimPatam »

michal wrote:The way I understand the neck model, you cannot lean without also tilting (unless you break your neck), thus the model can work out what is going on accurately. The only exception would be standing up or sitting down with your head craned back.

EDIT: Just to clarify, this is true only when your head is craned back at 180 degrees. Obviously you can lean without tilting otherwise.
OK i see what you mean, it's true that what when you lean/tilt your body your head also tilts with it, but it does at a much wider angle than when you tilt your head by moving just your neck; that happens when you face forward and when you face backwards 180 degree as well (by turning both your trunk and your neck in the same direction, no-one i know can turn his neck 180 degrees heheh).

I don't have time to make a sketch right now, which would be much clearer.. we'll probably have to agree to disagree on this one.


Anyway my point is: I think it would be good to have a more robust positional tracking solution that still works when you are not facing the camera, even for the average user, and i don't really like the idea of having to place and calibrate 2 separate cameras, that's the opposite of user-friendliness.
drifter wrote:The nice way would be to put a few leds on the back directly on the 2 straps.
Now the tricky part would be to get their positions relative to the front leds...
The only problem standing in the way of having LEDs visible on the back of the HMD is that pesky snowboarding elastic band (which i never really liked to begin with to be honest); using any other type of semi-rigid adjustment that could be easily measured by the HW itself would solve the issue.

Since the majority of HMDs out there don't actually use an elastic band i don't really see the point on keeping it when the benefits of removing it would be so great..
comham
One Eyed Hopeful
Posts: 16
Joined: Wed Feb 27, 2013 7:35 am

Re: POSITTRON: yet another proposal for positional head-trac

Post by comham »

1. For 360 degree tracking, extending the IRLED band round the whole head seems like it would be a lot cheaper than a second camera, although of course you couldn't put them on a flexible elastic tape thing because obviously you need the pattern of LED's to be constant.

2. Is it just me or does the placement of the LED's look a bit like the EURion constellation?
Image

Sorry, I try not to post much as I am a layperson and this forum has a high proportion of enthusiasts/professionals and I wouldn't want to dilute that
User avatar
drifter
Binocular Vision CONFIRMED!
Posts: 242
Joined: Tue Nov 27, 2012 9:39 am
Location: Little Britain

Re: POSITTRON: yet another proposal for positional head-trac

Post by drifter »

PatimPatam wrote:(...) Since the majority of HMDs out there don't actually use an elastic band i don't really see the point on keeping it when the benefits of removing it would be so great..
Interesting, btw what are the benefits of elastic straps ?
I see only 2 ones : it doesn't take place in the case, and the look is maybe more appealing for most people...
Are there other benefits ?
lossofmercy
One Eyed Hopeful
Posts: 45
Joined: Thu Oct 18, 2012 2:25 am

Re: POSITTRON: yet another proposal for positional head-trac

Post by lossofmercy »

If they go with a helmet, they better make it work well with small heads and giant heads (like mine). It would also limit headphone support.
User avatar
cadcoke5
Binocular Vision CONFIRMED!
Posts: 210
Joined: Mon May 24, 2010 8:43 pm
Location: near Lancaster, PA USA

Re: POSITTRON: yet another proposal for positional head-trac

Post by cadcoke5 »

You may already be aware of the retro-reflective technology, but in case you are not, I felt I should mention it. It would eliminate greatly simplify the creation of all those colored dots. Basically, a retro-reflector is what is in your bicycle reflector. You can also find self-adhesive vinyl. It is typically available in the auto parts stores, or even at Walmart. The only drawback that I can see to the retro-reflective stickers, is that they are flat, and therefore will not work over as wide an angle as your LED.

You need a light source near the observer (or camera) and the retro-reflective material will send a lot of the light back to the observer. So, it looks much brighter than the surrounding area.

To get a larger angle for visibility, you might paint small spheres, or even make globs of hot-melt glue.

For the white stripes in the road, they put down a special adhesive paint, and then sprinkle some retro-reflective glass beads onto it. One sources suggested using 80 grit sandblasting beads as a substituted to buying a 5-lb bag of the glass beads mail-order. I am also wondering if the white stuff they put onto Christmas cards, that look like snow, are those same sort of glass beads. They may have it in your local craft store. Carry a flashlight with you to test it by holding it next to your eye. If it is retro-reflective, it should show brighter than a white paper.

I just did and internet search, and was surprised to find "RUST OLEUM Reflective Paint" in a spray can. Krylon also makes a version. But, I have no idea how well either work.

For colors, you need something transparent, which means it cannot be a pigment. Jello is one solution for a cheap easy answer you may already have at your house. But I imagine any sort of clear overcoat method would work with food dye added, thought it would probably need to be water based to work with the water-based food dyes. Hair spray is one possibility.

-Joe Dunfee
User avatar
PatimPatam
Binocular Vision CONFIRMED!
Posts: 214
Joined: Thu Jun 28, 2012 1:31 pm
Location: Barcelona

Re: POSITTRON: yet another proposal for positional head-trac

Post by PatimPatam »

Thanks for the suggestion Joe, however i think i should point you to the 2nd page of this thread, where this was already discussed about a year ago:
http://www.mtbs3d.com/phpBB/viewtopic.p ... 2&start=44

I did actually have a quick test with reflective materials, but the main problem is that if they are flat then the reflective angle is very small compared to the viewing angle of certain LEDs (the ones i was using were almost 180 degrees for instance). Also the amount of reflected light can vary with the angle and with the lighting conditions of the room, so it's quite hard to make a robust recognition algorithm. The only way i can see it working would be using ball reflectors, but that's something i was explicitly trying to avoid with my design.


I must say in case it wasn't clear that i kind of abandoned this project a long time ago, basically other personal matters got in the way, and also most people seemed happy enough with the temporary Hydra solution for positional tracking, so no big incentive to go forward really. My main objective more than building an actual product was simply proving that it was a valid idea; after seeing the Crystal Cove prototype i believe i was not far off!

I guess with the Hydras being hard to find nowadays and the STEM and DK2 a few months away some people could still be interested in this, but i kind of lost interest myself.. Now that i have a little free time again i prefer to focus my efforts in building my own game; got some good ideas, but still in the process of evaluating different tools, engines, control schemes, etc.


Talking about that, my latest "crusade" is trying to convince the community / Oculus themselves that we need cheap optical motion controllers for PC, preferably if they were bundled with every Rift sold:
https://developer.oculusvr.com/forums/v ... =25&t=5912

Hopefully Oculus will announce something at GDC, but not letting my hopes too high about this!
User avatar
Fredz
Petrif-Eyed
Posts: 2255
Joined: Sat Jan 09, 2010 2:06 pm
Location: Perpignan, France
Contact:

Re: POSITTRON: yet another proposal for positional head-trac

Post by Fredz »

PatimPatam wrote:Thanks for the suggestion Joe, however i think i should point you to the 2nd page of this thread, where this was already discussed about a year ago:
The PS Move does seem to fit this goal quite well, I'm surprised not more devs used it in their demos. Wireless, very cheap, optical and with some buttons, that looks better than the Razer Hydra (expensive, wired, limited range) or the Sixense Stem (very expensive). At this point I don't think finger tracking is really required, that would be useful only in very specific game genres.
User avatar
PatimPatam
Binocular Vision CONFIRMED!
Posts: 214
Joined: Thu Jun 28, 2012 1:31 pm
Location: Barcelona

Re: POSITTRON: yet another proposal for positional head-trac

Post by PatimPatam »

Fredz wrote:
PatimPatam wrote:Talking about that, my latest "crusade" is trying to convince the community / Oculus themselves that we need cheap optical motion controllers for PC, preferably if they were bundled with every Rift sold:
https://developer.oculusvr.com/forums/v ... =25&t=5912
The PS Move does seem to fit this goal quite well, I'm surprised not more devs used it in their demos. Wireless, very cheap, optical and with some buttons, that looks better than the Razer Hydra (expensive, wired, limited range) or the Sixense Stem (very expensive). At this point I don't think finger tracking is really required, that would be useful only in very specific game genres.
Yeah i agree, i think the PS Move is pretty good. There are actually a few demos around that take advantage of it:
https://developer.oculusvr.com/forums/v ... =25&t=1687

I guess the main reason there aren't more is that it's not natively supported on Windows; you have to use "unofficial" APIs like these:
http://thp.io/2010/psmove/
http://code.google.com/p/moveonpc/


Having said that, it would be a bit silly to have to use 2 separate cameras (1 for the HMD + 1 for the controllers). And i don't really see PC gamers or Oculus supporting a Sony product (even if it's actually the best option available)..
WiredEarp
Golden Eyed Wiseman! (or woman!)
Posts: 1498
Joined: Fri Jul 08, 2011 11:47 pm

Re: POSITTRON: yet another proposal for positional head-trac

Post by WiredEarp »

Now that i have a little free time again i prefer to focus my efforts in building my own game; got some good ideas, but still in the process of evaluating different tools, engines, control schemes, etc.
What genre game are you working on PatimPatim?
User avatar
PatimPatam
Binocular Vision CONFIRMED!
Posts: 214
Joined: Thu Jun 28, 2012 1:31 pm
Location: Barcelona

Re: POSITTRON: yet another proposal for positional head-trac

Post by PatimPatam »

Hmm i would say it's a 1st person action adventure with a twist!

And not difficult to guess: the core mechanics require tight positional tracking and motion controls. :-P
Post Reply

Return to “VR/AR Research & Development”