Anyone witth RFID tech experience?

[Felsir]

I'm in a gaming group where we play some tabletop wargames. At the moment I have installed a webcam on the ceiling above our gaming table to make a snapshot each turn. I use the snapshots in MS Visio to overlay arrows and so on to produce battle reports.

So I was thinking; would it be possible to track the units on the table using RFID tags in their bases? I've read that it should be possible to read tags within a 3 meter distance? So I was thinking to use three readers (or 3 antenna) to triangulate the RFID tags. Unfortunately I have no idea where to look; the only readers I was able to find are either close contact readers or industry quality readers used in warehouses.
Anyone have experience or suggestions? Or feedback if my idea makes sense at all?

[SavannahLion]

I've read of systems that can activate and read multiple tags but AFAIK it doesn't triangulate positions but rather counts how many of a particular tag is within range.

I suppose you could triangulate the position of a particular tag but that's not an easy thing with off the shelf stuff.

Have you looked at cheaper options.  Fluorescent paints that can only be seen by the camera. No work rewuired to tue tsble but every piece must be painted with a dot visible to the camera. Or an IR or UV LED grid under the table. A computer can easily deduce where pieces move based on which LED goes on/off based on snapshots between moves. More work to modify the table but allows any piece to be introduced without modification as long as they adequately cover the LED. An LED sensor array will work too which would get rid of the camera.

[Howard_Casto]

I'm with SL on this one.  I looked into RFID for a similar project and discovered that it's original intended use (preventing shop lifting) is about all it's good for.  You can tell when a tag is within range of a scanner and the tag can hold some info, but other than that you can't really tell where the tag is, where it's pointing, ect.   The beauty and downfall of rfid is that the tags are powerless and thus their signal is very sporadic and weak.   

Not saying it can't be done, but your webcam approach is probably a better one.  Remember, webcams can see IR so that might be of some help. 

[Howard_Casto]

Oh... also look into AR cards, which is a much easier way to go.... all the big gaming companies seem to use this method.  As does the NFL btw.  All an ar card consists of is a easily recognizable pattern of lines and shapes that a computer can use to determine position.  For football they use the contrast of the green field with the white yard lines.  It's more than enough for them to get the position and the angle of the field to overlay graphics. 

[ark_ader]

Something like this?







It would need some calibration.  I'm sure you can get something like this working.


I am currently working on a HCI eye tracking project with NUIA (when it ships), so in theory the Kinect should be capable of tracking the whole table.

[Felsir]

You can tell when a tag is within range of a scanner and the tag can hold some info, but other than that you can't really tell where the tag is, where it's pointing, ect.

Yeah, I have been reading that it is possible to vary the range of each RFID antennae- thus being able to "sweep" from 100% to 0% and use that info as distance from tag to antennae- which in turn should be enough to determine the location. I doubt that it will be accurate enough for my purpose. Also so far I only found equipment from $1000 and up per reader so it will be way to expensive even to get a proof of concept going :-(

I'll see if I can tweak the webcam some more. The main problem with the webcam is that it is intended for use close to the monitor. I have a Microsoft Lifecam 5000HD, which does have an autofocus so the image is usable for my tracing purpose. I'll test if the AR cards (like the ones that come with the Nintendo 3DS right?) can be recognised.

SavannahLion, do you have a link to such paint? A google search resulted for me in paint which lights up in blacklights? Is that what you meant?

[SavannahLion]

I don't know. I think it's more like "invisible" paints. A while ago I was looking for some unique paints and that sort of thing kept coming up in my searches. Here is one such link: http://www.pfonline.com/articles/non-visible-infrared-ultraviolet-paint

I can never find the damn suppliers referenced though. As near as I can tell the military uses such paints to paint targets which are visible to cameras. In theory you really don't need a big blacklight per se.Maybe a few UV or IR emiting LEDs to excite the dye or paint and a camera without the uv or ir filters (or is it with? I can never get that part straight). There are other applications for that though as discussed in this thread: http://boards.straightdope.com/sdmb/archive/index.php/t-629405.html (ignore the alien crap). Here is a more practical application but expensive camera: http://www.sensorsinc.com/overview.html

I've always been interested in the technique and keep the idea filed away for experimentation but can never find the time to do it. On a side note, doesn't the Kinect leverage an IR camera?

[SavannahLion]

I almost forgot. If it helps check wiki for ultraviolet photography and infrared photography.

[kahlid74]

Earlier in my career I was a Network Engineer for a business who had manufacturing floors with 10,000 plus RFIDs.  They are way more powerful that just shoplifting but we're talking industrial RFID's.  RFID's were embedded everywhere so the line knew where a unit was and it's level of build.  You could find a unit anywhere in the factory based on where RFID's were.  The RFID's also held information relative to what was being built like Model/Serial Number, and what stage of the build it was on/had been completed.

So what you want to do is easily doable via RFID but for accuracy would take a large investment and some good coding.

As others have recommended you might want to even look into a few XBOX kinnect's and 3D model the board as you're playing.

[Howard_Casto]

Yeah I've seen similar setups but like he was saying, even an entry level scanner is 1000 bucks.  And that's to go with these cruddy supermarket rfid chips.  That's an awfully big investment considering what he's trying to do. 

[Felsir]

Yeah I've seen similar setups but like he was saying, even an entry level scanner is 1000 bucks.  And that's to go with these cruddy supermarket rfid chips.  That's an awfully big investment considering what he's trying to do.

Indeed, the investment is way too much to even get a prototype working. I was hoping that it would be cheaper to do (to compare, close proximity rfid readers can be bought for $20).