backtop


Print 47 comment(s) - last by frobizzle.. on Mar 25 at 1:39 PM


RFID Tags Printed on Plastic Foil  (Source: Gyou-Jin Cho/Sunchon National University)
Printing process uses carbon nanotubes

RFID tags are in many of the products that we already buy today and the promise of RFID in the future is that we may not even have to stop at the register to checkout at the store. In the future, with prolific RFID tags more powerful than what we have today, all we might need to do is walk out the door with our carts and our total would be computed automatically. Today's RFID technology, however, is prone to hacking, which was demonstrated when researchers were able to clone an RFID passport while driving by it.

Before we can get to the point where store inventories are able to be done in real-time using RFID tags, we need to have cheaper and more efficient methods of producing the tags and the tags need to hold more information and use less power. Researchers at Rice University and Sunchon National University in Korean are working on a joint project using RFID tags that are printed on a roll-to-roll process that uses inks embedded with carbon nanotubes.

The printing process is able to make RFID tags continuously with a cost of pennies each. The technology for the ink was first invented in the Rice lab of James Tour and was at the time used to print thin-film transistors, which are a key part of RFID tags printed on paper or plastic. Gyou-jin Cho from the Sunchon National University in Korea says that professor Tour was the person who recommended using single wall nanotubes in the ink.

Cho said, "Professor Tour first recommended we use single-walled carbon nanotubes for printing thin-film transistors." Tour says that Rice owns half the patent that is pending on the technology and states, "Gyou-jin has carried the brunt of this, and it's his sole project. We are advisers and we still send him the raw materials." Tour's lab is where the carbon nanotubes needed for the ink are produced.

Most RFID tags that are in use today are made from silicon-based materials. Paper or plastic tags that can be printed would dramatically reduce the cost of making RFID tags. Printed tags could also replace the bar codes that are printed on all packages now.

The process developed by the researcher is able to print one-bit RFID tags complete with an antenna, electrodes, and dielectric layers on a roll of plastic foil. The process still needs to be refined and improved before it is practical. Cho is working on 16-bit tags that could hold enough data to be useful in real world applications and still be printable on paper.

The printed RFID tags are passive and need no power source, which is what makes them so cheap to produce. The tags only give off the data on them when hit with radio waves at the correct frequency. The RFID tags also have to be reduced to about a third of their current size to be printable on packages.

The team is also working on increasing the range that the tags can be read from. Currently the printed tags can only be read from a very close distance to the transmitter. To be useful in inventorying an entire store or warehouse the tags need a range of about 300 meters. 

"Right now, the emitter has to be pretty close to the tags, but it's getting farther all the time," said Tour. "The practical distance to have it ring up all the items in your shopping cart is a meter. But the ultimate would be to signal and get immediate response back from every item in your store – what's on the shelves, their dates, everything. At 300 meters, you're set – you have real-time information on every item in a warehouse. If something falls behind a shelf, you know about it. If a product is about to expire, you know to move it to the front – or to the bargain bin."



Comments     Threshold


This article is over a month old, voting and posting comments is disabled

RE: How's that gonna work?
By geddarkstorm on 3/19/2010 3:11:19 PM , Rating: 1
It's called a Fourier Transform http://en.wikipedia.org/wiki/Fourier_transform

We do this in NMR while collecting data to determine the structure of macromolecules three dimensionally. To parse the atomic frequency data, one has to deconvolute millions of frequencies all squished together in one giant signal. That's what a fourier transform takes care of, and can be done in real time.


RE: How's that gonna work?
By porkpie on 3/19/2010 5:44:02 PM , Rating: 1
No. A FFT just transforms a signal from the time to frequency domain or vice versa. If you have one signal of a certain frequency buried in noise of million different frequencies, a fourier transform can extract it...but if you have several thousand transmitters all on the same frequency, it's not going to help.

Some combination of discrete time signal processing and adaptive beamforming on the receiver end might work in theory, though...essentially taking advantage of the fact that each tag has its own unique position in 3D space.


RE: How's that gonna work?
By jerkweed on 3/19/2010 11:12:01 PM , Rating: 2
I think you've got 2 misconceptions here. First, every single item on the shelf will have a unique ID...it's not like barcodes where every SKU is the same code. Each individual package has it's own completely unique ID (which means they can track exactly which can you bought). Second, tag collision is basically a solved problem...they can just use that same completely unique identifier as the code to delay reply by ex-microseconds.


RE: How's that gonna work?
By porkpie on 3/20/2010 12:56:46 AM , Rating: 2
"First, every single item on the shelf will have a unique ID."

A unique ID does not imply a unique transmission frequency, nor does it help in any way with collision issues.

"Second, tag collision is basically a solved problem."

Stochastic retransmittal is useful in cases of a few dozen tags; in the case of hundreds of thousands of tags energized simultaneously, it's nearly useless, and can even make the situation worse, which a little reflection should convince you.

RFID tag collision is not a 'solved problem'; its actually a very hot topic in the literature at present.


RE: How's that gonna work?
By AssBall on 3/20/2010 3:17:53 AM , Rating: 2
why would you have several thousand tags at the same frequency? If it is a warehouse give each product a slightly different frequency, and if its a store, give each product a slightly different frequency.

There is still development to be done on this, to be sure, but there are some simple things that can be done to avoid collision issues.


RE: How's that gonna work?
By AssBall on 3/20/2010 6:16:42 AM , Rating: 2
Oops, that should read "pallet" and "case". I still don't see how this would be difficult, once the RFIDs could be produced cheaply. The whole system would need some major tweaking and money to setup and install, but I can see how it would work.


RE: How's that gonna work?
By porkpie on 3/20/2010 11:21:11 AM , Rating: 2
" If it is a warehouse give each product a slightly different frequency"

Then how do you read them? A reader that can energize, sweep, detect, and lock onto any of 100,000 different frequencies in real time is not trivial.

Some RFID collision avoidance schemes do incorporate a certain amount of spectrum expansion. But its just not practical to give each tag its own frequency in the general case.


RE: How's that gonna work?
By Solandri on 3/21/2010 5:37:14 AM , Rating: 2
quote:
No. A FFT just transforms a signal from the time to frequency domain or vice versa. If you have one signal of a certain frequency buried in noise of million different frequencies, a fourier transform can extract it...but if you have several thousand transmitters all on the same frequency, it's not going to help.

Some combination of discrete time signal processing and adaptive beamforming on the receiver end might work in theory, though...essentially taking advantage of the fact that each tag has its own unique position in 3D space.

Actually, he has a good point. A FT will also do spatial convolution/deconvolution. e.g. If you have the amplitude and phase information for light passing through a window in a certain direction, a FT applied to it will act like a lens, and "focus" the light to create an image of what the scene looks like through that window. I hadn't considered tackling the problem in the spatial domain (since there is no concept of a spatial domain in computer network communications).

http://sharp.bu.edu/~slehar/fourier/fourier.html#a...

With that mental barrier lifted, I can think of yet another way to solve this problem: Use two or (preferably) more receivers. Then you can use tomography to figure out the locations of each transmission, thus separating out each transmission. We did something similar with underwater acoustics (using both tomography and the phase of the arriving signal to distinguish between difference sources).

But those were just a handful of signal sources and the tomography computations took several hours back in the mid 1990s. I'm not sure how viable this would be with tens or hundreds of thousands of RFID tags. Maybe treating it more like phased array radar would simplify the whole thing. With a phased array receiver, you could "aim" it at a certain spot in the store, essentially picking up only the RFID responses form that location. Since the aiming point is determined mathematically (no moving parts), you could rapidly scan through the entire store, "listening" for responses one by one.


RE: How's that gonna work?
By porkpie on 3/21/2010 9:04:19 AM , Rating: 2
quote:
Maybe treating it more like phased array radar would simplify the whole thing. With a phased array receiver, you could "aim" it at a certain spot in the store
That's what I suggested above, using adaptive beamforming and discrete time signal processing to locate tags spatially.


"If you mod me down, I will become more insightful than you can possibly imagine." -- Slashdot














botimage
Copyright 2014 DailyTech LLC. - RSS Feed | Advertise | About Us | Ethics | FAQ | Terms, Conditions & Privacy Information | Kristopher Kubicki