RFID: Old technology, new possibilities

12 billion-mile range
Theoretically, RFID tags can operate on any frequency but in practice they are limited to eight bands, one or more of which are allocated by national regulators around the world. These are often further classified into low, intermediate and high bands: just as with broadcast radio on medium wave, short wave and VHF, the bands have different characteristics suited to different tasks.

The low band, between 100 and 500 kHz, is suited to short and medium range use; it has good penetrative properties for buildings and can frequently work around conductive barriers, liquids and other materials that can shield higher frequencies. It's got a low reading speed so can't transfer much data, and is commonly used for access control, livestock identification and inventory control.

The intermediate band, between 10 and 15 MHz, is also suitable for short and medium range. It can be made cheaper than low band RFID, isn't quite as robust in difficult conditions and has a medium reading speed -- smart cards and access control often use it.

High band is UHF and microwave, with frequencies at around 430 MHz, between 800 MHz and 1 GHz, and on unlicensed bands such as 2.4 and 5.8 GHz. Given line of sight between the reader and the tag, it can achieve long range but is easily thwarted by obstructions, walls and so on. It is also potentially prone to problems with reflection and interference. Until now, it has been the most expensive form of RFID, but because components and antenna for high band systems are much smaller than the alternatives, they lend themselves to mass production and high levels of integration. High band is capable of much higher read speeds than the others, which has made it popular for RFID tasks involving inventories of containers and palettes, or where encryption is required: the American regulator, the FCC, has recently allocated a high band at 430 MHz where tags can send back messages of up to a minute in length.

There are two main protocols currently in use, Class 0 and Class 1, which are incompatible. A new standard, EPC UHF Class 1 Generation 2, is intended to unify these, tighten up security and create a higher read speed. Operating in a high band frequency range, the proposed standard has a 96-bit ID, a 32-bit kill command that silences the tag for good -- an important feature for privacy concerns -- and systems using it will be able to manage 1700 tags a second in the US and 600 a second in Europe -- regulatory differences between the two regions put different bandwidth and power restrictions on readers.

RFID's future is bright, interesting and complex. Although the breakthrough price for an RFID tag is commonly felt to be five cents, a cost low enough to supplant barcodes in a very wide range of items, this will require considerable work in producing and packaging very small silicon chips, around 0.1mm square, in quantities of billions. Companies such as Alien Technologies and Matrics are concentrating on this approach. Another technology with great promise, more exotic but capable of even cheaper tags, is printing them directly onto packaging using conductive and semi-conductive inks. Xerox has recently announced it has developed a method of ink-jet printing circuits using polythiophene inks which can combine conductors, insulators and active devices -- the three classes of component necessary to build a tag -- without needing special atmospheres during printing. Estimates for the first practical printed tag range from three years to 15 years, although gigahertz parts are going to take longer to develop than slower, HF or UHF circuits.

When this happens, RFID tags will be only fractionally more expensive than barcodes -- potentially, both could be printed at the same time. At that point, RFID will become ubiquitous and the first machine-readable world will be on its way.