Specs of VR Unit....

[T3rra]

Okay, I'm no genius, so if something doesn't work out, well you'll know why lol....
Anyways...

Okay, so I've been thinking up a VR unit, not working, or not deep enough, on an HMD, but a full body unit; 2 for legs, 2 for arms, an HMD, and a controller, plus some other upgrades, but thats not till later,...

The parts I would need are 6dofs for each unit - 6DOF - I did NOT make this, this is just a guy on a forum who did this.... the websites there...

I have developed a 6DOF for < $35 using Wii nunchuck and Wii motion plus. Works better than the Sparkfun 6DOF
(http://aeroquad.com/showthread.php?12-L ... is-project)

anyways, unless there is a cheaper than 35$ alternative, I'm going to go with this one.
Next is a gps receiver....

http://www.ecrater.com/p/13585666/eb-36 ... board-modu le-with?gps=1
EB-3631 GPS Engine Board Module with SiRF Star III Chipset
Price: $23.98
Dimensions: 0.63 in x 0.51 in x 0.08 in (1.6 cm x 1.3 cm x 0.2 cm)
Weight: 0.11 oz (3 g)

note, I don't know much about gps receivers, I just went with what looked like it made sense....
This is all more in depth, if you have questions just let me know.

Next is the transmitter - there are two types - 1 for the AP's and LP's (arm pads and leg pads)which sends a signal to the HMD, which has a video/audio receiver - I actually have a question about this, perhaps you guys can help me out; is there a receiving chip which receives up to 10 different signals (basic code, like positioning on a map, or triggers) is this possible? anyone know where I can get one for cheap?

Transmitter – Wireless RCA Video Audio Transmitter and Receiver, 2.4GHz 1W
Today: $24.99 ($24.99 each)
http://www.meritline.com/wireless-rca-v ... 4AodA0oAcQ

Features & Specifications:
•Operating Frequency: 2414 – 2468 MHz (2.4GHz)
•Supports four selectable channels (1-4)
•Supports multiple receivers, ISM Band from 2414~2468 MHz
•Supports NTSC/PAL Video Standards.
•Range up to 100 feet indoors (Results will vary based on how many walls, wall thickness, and other conditions)

I don't know if within an enclosed location - possibly cover the walls with a type of aluminum to block out other signals (or not) so there is no clutter of signals no lag, so basically the HMD has 2 receivers or 1 if possible, one for video audio and the other for data signals...

I don't know much about motherboard chips, still haven't looked into it yet...
a usb slot on the HMD, create a program which allows you to set which signals the HMD picks up ( only picks up specific signals from APs and LPs, rejects other signals (poss?)

A battery - the Itouch one - BATTERY (HMD) - iPod Touch 2nd Gen and 3rd Gen Battery Our Price: $26.74: $16.99 On Sale

I haven't looked into the HMD one yet and how much power will be consumed by the lenses.... can't be too big tho obviously....

the lenses like I mentioned before are 5.6mm lenses, and palmers right, I don't know anything about lens warping (or whatever he said )

I mean thats pretty much it....
I figured the cost of these units would in total be like 700$ but that cost is covered if you charge players 10$/hour and give them perks like Atlantic city does for its players - shirts, custom controllers, and player cards (if you get a community to agree with you, you can create cards which can be used for discounts at those locations; food, clothing, plastic weapons for in game - it's kinda like a theme park)

I made a sheet about how much you could expect to pull in if you had 16 units in a small paintball esque area, with a small margin of error (prob have to change that, and I haven't finished it either, have to take some other factors into consideration) but i'll do that when I get the chance.

Like I said, I'm not a genius, so please don't over criticize, besides, it's not right to do that :/ try to be more pleasant? (I'm aware, I'm not your mother, but still.......)

I dunno, I left out some stuff, but if you have a question i'll see if I have an answer

[Namielus]

Even if you use one transmitter per eye that PAL/NTSC transmitter is nowhere near good enough for the resolution and fps you need for a HMD

[brantlew]

A few critiques. If you are envisioning a system that tracks character position as well as accurate body poses, then the devices in your list will not be adequate to accomplish that.

First of all, GPS accuracy even under perfect conditions is only about 1 meter. And that's outdoors in a flat treeless area away from any buildings with line of sight to the entire sky. Anything less than those conditions and your accuracy will range from 1 to 10 meters and move a good bit. And indoors, GPS is too weak to penetrate so even when you get a signal your accuracy will be about a 100 meters or more. So GPS is inadequate for accurate position tracking.

Secondly, those 6DOF sensors don't work the way you think they do. They will track changes in orientation fairly well so you will know what angle they are pointing, but you will have no idea what location they at in relation to each other. If you move your leg, you can track the bending of it but the data might show that it flies off your body at 10 miles per hour. You would need a lot more sensors and/or a different type of sensor to accomplish anything like skeletal tracking.

It's good that you are starting to discuss this more openly, because there is a mountain of knowledge on this forum that you would need to digest (in a general and non-technical way) before you would be able to take on an undertaking like this. If this type of installation were easy and cheap to accomplish there are about 50 people on the forum that would have done it already. But the devil is in the details.

[PalmerTech]

A short list of issues:

1) That sensor is not suitable for limb tracking.

2) The GPS is not fast enough for positioning.

3) Those video transmitters have the same limitations of most composite video transmitters: max res of 640x480 in 2D, lots of latency, and no way to have many transmitter/receiver pairs operating in the same place. They eat a lot of power, too.

4) Your battery would power a Rift for perhaps half an hour. It would probably not even handle the current draw of all your other stuff, but assuming it could, my rough estimates show a battery life of about 9 minutes.

5) Lenses do not consume power. It does not matter how big they are.

6) A 5.6mm lens is about the size of a pea. I don't think it would work very well.

7) When you make your money sheet, you have to factor in employee costs, liability insurance, leasing the land needed, taxes, equipment repairs etc. Running a physical establishment like a laser tag or paintball facility is expensive.

[Namielus]

I thought the 5.6mm referred to focal length, since I have seen one for dslr.
But that still would not make much sense.

[RoadKillGrill]

First thing you should do with this is ignore cost, once you have a functioning demo it is easier to calculate true cost and figure out what is necessary and what can be adjusted.
Even if the suits cost 10k a piece they could make they money back rather quickly. Venues are much more willing to spend a lot of money on a system if it will attract people to their place. Arcade machines can cost more than cars but they will make money.

For a cheap wireless plug and play HMD the fatshark an option, it wont be as good as a rift but is already assembled and wireless. It designed for hobby planes but the unit works ok for VR. I used it in College for VR work.

I think brantlew got most the reasons that GPS will not work (its slow too), there is still a many options available for local accurate tracking in a known environment.

6dof sensors suffer from drift, after a few minutes you would need the player to stop and pose to get the tracking semi accurate. 9dof sensors can preform this task better, they will be consistent for hours.
The inertial system I made used 17 for full body capture, if the feet, hands and shoulders were removed it can be 10 IMUs.
The minimum count of IMUs to get a rough body capture is 7, 1 for each foot, 2 per each arm, 1 chest and 1 head. With the minimum amount the legs will be quite fun to program to look right but I think should be possible. You would use inverse kinematics to solve the knee position, since the feet at typically on the ground figuring out placement and constraints is a little easier. The arms are going to be closer to the user so error will be more noticeable and lack a normal resting position so I don't think you can remove 2 IMUS from the arms for tracking. Chest and head I think are obvious on their use, chest might be able to be estimated by the upper arms and head orientation but I would need to try that. The less sensors the more data will need to be estimated.


Ignoring all the hardware necessary right now, have you ever designed a game before? Even as simple as a board game? If you haven't then the hardware is the least of your worries.
The biggest thing missing from this list is software. When it comes down to what is going to make this whole thing work or not is going to be the software. Fusing all the data together to get a player position, pose, and interaction is not an easy task. I would guess when doing this project significantly more man hours will be in software than in the hardware. This isn't even including the GAME section of this project. Even with a complete suit and HMD solution it could take months to make it into a game that was playable.

[WiredEarp]

@ T3rra: as Brantlew, Palmer, etc, have pointed out, one of the main issues is going to be the GPS. If you look around these forums, you'll see another thread about almost exactly what you are talking about (VR Lasertag), in which we discuss all the different tracking technologies available currently.

GPS will not cut it in an indoor environment usually. Optical is currently looking to be the most promising tracking method for tracking multiple people within a large indoor space. While there are apparently professional optical optical systems that can do this already, they cost a hell of a lot. Basically, if you can solve the tracking issue, then most of the other parts of VR lasertag be solved with varying amounts of effort. Doing something like that and keeping the tech to yourself as long as possible would be one way of getting a leg up on the competition in this field. Otherwise, you will be competing against everyone else who has similar ideas and is also waiting for the right tech to emerge to put it into reality. That is not to say the VR lasertag market wouldn't be big enough for multiple main vendors to have competing products, I think it would be a huge market, just looking at the amount of people who play lasertag and paintball.

[T3rra]

When I said GPS tracking, I meant an in game/hardware based tracking method. For example; map placement is based on an in game layout. Tracking is based on x,y,z axis within the game, not a satellite based tracker..... unless I misunderstood that lol....

as for the pal/ntsc, what would you suggest I replace it with? Should I go with palmers advice and stick to the rift specs?
I mean I'm an all around kinda guy when it comes to this, not developing only an HMD you know? whatever works I'll look at the rift.

as for skeletal tracking..... why would the sensors detect your body traveling at x miles an hour? isn't there a way to change that through the software? I see the vids for head tracking, isn't it the same for any other part of your body? I mean the head isn't going a thousand meters/sec...... I dunno, what am I not getting? it has to be the software.... right?

as for 5.6mm, I just looked up some stuff on alibaba.com I guess I shoulda looked into it more.... still don't understand the technical sides of it, confusing tbh, but it's not so bad lol... god I feel like a child ..... more rift specs!

Hmm... roadkill, is there something I'm missing? I mean if you give the character body physics based on our gravity, shouldn't the body parts move in relation to how you're moving around? what if you assigned movements to body parts in game/maya/3dsmax? I mean you give it proper physics dimensions and it's arms won't go flying about right? or is it not that easy i'm going to guess? I'm sure there are people working on this, like you for example....

And I dunno, I don't wanna go into software, I'll just leave that to the pros, and i'll just look for specs that I need. Good to know I was off with almost/everything I posted =_= lol....

what do you guys suggest I do next minus more research? :/ thanks for the input!

[T3rra]

@ T3rra: as Brantlew, Palmer, etc, have pointed out, one of the main issues is going to be the GPS. If you look around these forums, you'll see another thread about almost exactly what you are talking about (VR Lasertag), in which we discuss all the different tracking technologies available currently.

GPS will not cut it in an indoor environment usually. Optical is currently looking to be the most promising tracking method for tracking multiple people within a large indoor space. While there are apparently professional optical optical systems that can do this already, they cost a hell of a lot. Basically, if you can solve the tracking issue, then most of the other parts of VR lasertag be solved with varying amounts of effort. Doing something like that and keeping the tech to yourself as long as possible would be one way of getting a leg up on the competition in this field. Otherwise, you will be competing against everyone else who has similar ideas and is also waiting for the right tech to emerge to put it into reality. That is not to say the VR lasertag market wouldn't be big enough for multiple main vendors to have competing products, I think it would be a huge market, just looking at the amount of people who play lasertag and paintball.

My idea is as follows; see from what I gather, gps is just like 4+ sats that transmit a signal and using the meet points of the waves, they send you your exact position right? So what if you used 4 baby sats (not satellites, like 4 "pads" that are installed on top of a green box. - The player steps in, then the arm pads and leg pads receive these signals, and transmit it to the HMD, which then transmits to a "tower" 1 - 200 ft away, it then turns relays back your position, and just once too, cause afterwards you'll be moving in game, you won't need a tracker anymore.... right? I mean if you have 6dof on all your parts, you should be mapped in game.....right? uhoh i'm getting nervous now lol.... and it's just that, 4 pads on an elevated area, maps your points, relays to the tower, you get your position in return. won't be needed afterwards, so it should be small and cheap right?

[brantlew]

as for skeletal tracking..... why would the sensors detect your body traveling at x miles an hour? isn't there a way to change that through the software? I see the vids for head tracking, isn't it the same for any other part of your body? I mean the head isn't going a thousand meters/sec...... I dunno, what am I not getting? it has to be the software.... right?

I would suggest starting with this video about inertial sensors as a primer on the topic. If you want to hear the reason why body parts would fly off at x miles per hour - you can skip to the 23 minute mark.

http://www.youtube.com/watch?v=C7JQ7Rpwn2k

[T3rra]

as for skeletal tracking..... why would the sensors detect your body traveling at x miles an hour? isn't there a way to change that through the software? I see the vids for head tracking, isn't it the same for any other part of your body? I mean the head isn't going a thousand meters/sec...... I dunno, what am I not getting? it has to be the software.... right?

I would suggest starting with this video about inertial sensors as a primer on the topic. If you want to hear the reason why body parts would fly off at x miles per hour - you can skip to the 23 minute mark.

http://www.youtube.com/watch?v=C7JQ7Rpwn2k

I don't understand half of what he's saying
could you break it down into laymans terms?

[PalmerTech]

My idea is as follows; see from what I gather, gps is just like 4+ sats that transmit a signal and using the meet points of the waves, they send you your exact position right? So what if you used 4 baby sats (not satellites, like 4 "pads" that are installed on top of a green box. - The player steps in, then the arm pads and leg pads receive these signals, and transmit it to the HMD, which then transmits to a "tower" 1 - 200 ft away, it then turns relays back your position, and just once too, cause afterwards you'll be moving in game, you won't need a tracker anymore.... right? I mean if you have 6dof on all your parts, you should be mapped in game.....right? uhoh i'm getting nervous now lol.... and it's just that, 4 pads on an elevated area, maps your points, relays to the tower, you get your position in return. won't be needed afterwards, so it should be small and cheap right?

The general idea you have described (Base stations transmitting or receiving a signal of some kind, and mounting the other end on the player) is not a new idea, or simple and cheap. If it was, it definitely would have been done already. Look into magnetic tracking, it is the closest to what you seem to have in mind, but it has pretty severe limitations. Making a tracker is not as simple as you think it is, it is a multi-year effort for a talented team of dedicated professionals.

[brantlew]

I don't understand half of what he's saying
could you break it down into laymans terms?

Inertial trackers are good for angles and general motion direction. They suck horribly for specific X,Y,Z coordinate positions.

[FingerFlinger]

To clarify how GPS works, it is a time-of-flight system, meaning that the GPS device measures the amount of time it takes for a signal to travel from a satellite to the device. The distance to the satellite is proportional the travel time of this signal. Communicating with at least 3 satellites allows the device to position itself in 3-space w/ respect to these satellites.

So, you are talking about using a time-of-flight RF system for your tracking. The problem with using these terrestrially is that the RF signals move at the speed of light, and there simply isn't much hardware available that can accurately measure the extremely short of amount of time it takes for these signals to travel 200 feet, as you have spec-ed out. Krenzo made a thread about his RF tracking system here, which is based on angle-of-arrival. I know a little bit about the various Real-Time Location System solutions out there, and as far as I know, he's the first one to crack this problem (of measuring very short RF phase-delays/times-of-flight to a high accuracy). And he did it through some very impressive and hardcore engineering.

I don't understand why you think that you don't need the tracker once you are "in the game". If you are just using it for initialization, then why use one at all? Also, you definitely need one, especially in an enclosed environment; pure dead-reckoning solutions will have you running into walls like crazy. As noted above, optical is probably the most effective way to go right now.

EDIT: @ brantlew That's a great video; I love Google Tech Talks!

[Zaptruder]

"Don't over criticize"?

If you want to make it in this world, you're going to need thicker skin then that.

[RoadKillGrill]

Hmm... roadkill, is there something I'm missing? I mean if you give the character body physics based on our gravity, shouldn't the body parts move in relation to how you're moving around? what if you assigned movements to body parts in game/maya/3dsmax? I mean you give it proper physics dimensions and it's arms won't go flying about right? or is it not that easy i'm going to guess? I'm sure there are people working on this, like you for example....

And I dunno, I don't wanna go into software, I'll just leave that to the pros, and i'll just look for specs that I need. Good to know I was off with almost/everything I posted =_= lol....

Giving it gravity doesn't just make it work, the IMUs are great at orientation but fair to poor for positioning. Using some assumptions on the constraints of leg motions and a bit of IK the feet might work. The arms don't have a much resting pose so it would be harder to guess positioning. Adding trackers to the upper and lower arms gives enough data to set the orientation and gives the effect of accurate positioning. The length of the bones on the virtual model need to be proportional to the real model for this method to work.

What will making a shopping cart of parts do without software? I have easy access to all the parts you listed and then some but I am not even close to making a multiplayer full suit AR/VR game. I even have a location large enough for this already picked out but I don't have the time to write the rest of the functional code then game code on top of that. In addition the level needs to be made and some artwork done. A to scale virtual world will take time to make, even if its just blocks.

Learning how to do some software is an important step in making stuff. Don't leave it to the "pros", start figuring out how to make software!!!! Its also pretty much free to write software, if it doesn't work only time is lost. If you have some those IMUs already start trying to do things with them, make it into a mouse, do some gesture recognition, make an alarm system, make a game that uses a non traditional input system, if you have two try tracking arm movements, head movements, anything!!!!!!

[rmcclelland]

I've thought of this tracking idea also, mostly for my work (space tech development). Basically, a highly localized GPS. The best currently available system (which we have used) is Differential GPS. Google it.

I believe part of the magic of GPS is the extremely accurate clocks, which are also extremely expensive (though getting cheaper). GPS timing takes relativistic effects in to account (time passes differently further from the earth and also depends on the GPS spacecraft speed).

An inexpensive tracking system were localized base stations track a moving point to mm accuracy would be useful in many industries.

[Krenzo]

I believe part of the magic of GPS is the extremely accurate clocks, which are also extremely expensive (though getting cheaper). GPS timing takes relativistic effects in to account (time passes differently further from the earth and also depends on the GPS spacecraft speed).

GPS satellites only need accurate clocks because they're designed to operate for a long time before requiring clock corrections from base stations on Earth. If frequent clock corrections aren't an issue, they could run with cheaper clocks. Atomic clocks (used by GPS) are good at long term stability, but there are better clocks for short term stability. There are devices available that use a local ovenized oscillator which is much better at short term stability and communicate with GPS to achieve atomic clock timing for long term stability. If you want a system that is only going to run for a few hours, you don't need to worry about how many seconds you're going to be off after running for a year.

You can't look to GPS when talking about a localized positioning system. GPS has different design priorities.