Sponsors of Wiimoteproject.com
IR Pens for Wiimote Whiteboard
Pages: [1]
  Print  
Author Topic: Rapid X, Y, Z detection?  (Read 5442 times)
0 Members and 2 Guests are viewing this topic.
*
Karma: +0/-0
Posts: 13
Offline Offline
View Profile Email
« on: August 06, 2008, 05:01:16 AM »

I'm a programmer / game developer, but I don't know much about hardware. I'd like to know what sort of device would be best for continuous (at least 30hz) positions relative to a sensory origin. I'm talking about several (four or more) wireless devices passing very small data packets (XYZ coordinates with double precision means 24-bytes per frame, per device), presumably via WiFi. To hell with the accelerometer or buttons. The main shortcomings, as with most consumer technology, are price and size: devices similar to what I describe are a bit too expensive (ideally under $10 each), and a bit too large (ideally smaller than matchbooks).

The Wiimote, unfortunately, does not satisfy these needs. It communicates its position relative to the screen via infrared detection, and must be pointed at the screen to work properly.

My goal is a difficult one, and it sounds like it's from a space opera or comic book. I want to attach several of these nodes to someone's body as a form of HIDs - really advanced/immersing gameplay, if achieved. In fact, nothing short of breaking new ground in the video game industry.

I know the technology do do exactly this has been around for at least a decade; however, it has only been used in academic contexts - expensive and nonstandard. Now, finding a homebrew means of achieving such an input method is likely within reach. I have come to this project in my zeal. You guys have the knowledge and, likely, the resources. Can this be done?
« Last Edit: August 06, 2008, 05:05:28 AM by Ringer » Logged
*
Karma: +0/-0
Posts: 6
Offline Offline
View Profile
« Reply #1 on: August 22, 2008, 04:16:02 PM »

You need at least three points of reference that aren't co-linear, i.e. they form a triangle not a line. From this you can work out a position in 3D space.

 Although for each reading of the three reference points there are actually 2 resulting possible 3D positions, one of them can be excluded by using common sense. For instance GPS satellites assume you are on the ground and not floating in space.
Logged
Support Admin
*
Karma: +11/-3
Posts: 566
Offline Offline
View Profile WWW
« Reply #2 on: August 22, 2008, 05:11:37 PM »

...my GPS also displays meters above sea level (altitude)


//UndCon
Logged

*
Karma: +0/-0
Posts: 6
Offline Offline
View Profile
« Reply #3 on: August 22, 2008, 06:32:48 PM »

...my GPS also displays meters above sea level (altitude)

//UndCon

 If you'd like to take a GPS reading at around 40,000 Km above sea level you'll get my point.
Logged
*
Karma: +0/-0
Posts: 13
Offline Offline
View Profile Email
« Reply #4 on: August 26, 2008, 03:17:25 AM »

I'll try to implement sea level in my prototype.
Logged
Support Admin
*
Karma: +11/-3
Posts: 566
Offline Offline
View Profile WWW
« Reply #5 on: August 26, 2008, 09:46:12 AM »

If your software is assuming you are on the ground, then it is not intended to be used outwards in space.

As most of us don't have the ability to go beyond 40.000Km they don't implement that code into the software but a simple if clause is enough to determine if this is true or false.

No big deal...


Back to the topic...

I know someone modified the wiimote to be faster as the hardware is capable of doing so...

Mocap isnt very cheap and requires i'm afraid...
Logged

*
Karma: +0/-0
Posts: 13
Offline Offline
View Profile Email
« Reply #6 on: August 26, 2008, 01:25:09 PM »

I've pretty much turned over this method (attaching hardware to the body) in favor of regular Wiimote altercations. I think infrared lights could be the key to better detection after all. If IR LEDs are on the body (knees and hands, I guess) and the Wiimote is aimed the waist, there's a good chance it could track everything. I read that the Wiimote's camera can track four points at a time... But can it track them accurately and distinguish them, or does it truncate their positions to two points? And what's the lens angle of this thing?

Of course, I'll run into testing this soon enough, but if you already know how the camera sends point data (especially if four coordinate pairs are passed), you can give me a very quick answer!

If the Wiimote isn't naturally up to the task, I'll have to (read: get to) use my own hardware for detection, like these guys. And as the whole community seems to, I encourage you to try it yourself! Most of my "free" time gets tied up otherwise, so as much as I want to see this turn out, I might not be able to build a dozen experiments.
Logged
Support Admin
*
Karma: +11/-3
Posts: 566
Offline Offline
View Profile WWW
« Reply #7 on: August 26, 2008, 03:32:32 PM »

I havent encountered any problems with the wiimote using 2 IR-pens

BUT...The wiimote cannot tell if it is source 1 or source 2 that is lit...

Lets say you are tracking 3 dots

hip, knee, ankle

and then you suddenly drop 1 of them - knee...

there is nothing to tell them apart (besides position and movements) so you might have a corrupted model...
Logged

*
Karma: +0/-0
Posts: 6
Offline Offline
View Profile
« Reply #8 on: August 27, 2008, 05:14:58 PM »

You wouldn't be able to tell if the person was doing a handstand or not.
Logged
Support Admin
*
Karma: +11/-3
Posts: 566
Offline Offline
View Profile WWW
« Reply #9 on: August 27, 2008, 06:48:58 PM »

exactly - there is no ID on the IR-dots


//UndCon
Logged

Pages: [1]
  Print  
 
Jump to:  

Clicky Web Analytics