April 27, 2005

Companies Find Workarounds for RFID Roadblocks

By Evan Schuman, Ziff Davis Internet

Retailers and consumer goods manufacturers experimenting with RFID are still seeing higher-than-expected error rates, but they are starting to work out low-tech workarounds, according to a major European RFID analyst company that has analyzed about 1,400 recent radio-frequency identification trials.

The most common problem has been reader interference coming from either metal or fluid, which was a major problem for The Gillette Co. in trying to scan its razor blades, said Raghu Das, a managing director at IDTechEx Ltd., which created the report.

Even Philip Morris International Inc. found that the thin metal on the paper wrapping it uses to surround certain cigarettes was causing problems, Das said.

He added that the tobacco giant admitted that the paper—perceived as enhancing freshness—didn't serve any purpose, but that it would have been too expensive to redo the packaging.

Kleenex-maker Kimberly-Clark Corp. accidentally discovered a simple workaround, Das said. During normal assembly-line operations, all pallets are fully shrink-wrapped. The machinery that does this slowly spins the pallet around.

The company discovered that an RFID scan taken during that rotation delivered much more accurate reads, Das said, apparently because it allowed the reader to "see" the tags from different angles and grab the best view.

But testing also has raised the possibility that putting more distance between the tag and the interfering material by using certain kinds of shrink-wrapping or substrates (surfaces where labels/tags can be printed or placed) could play a role in reducing errors, Das said.

Based on the RFID trials that IDTechEx was able to identify, the United States is still overwhelmingly dominant, attributable mostly to Wal-Mart. Of all of the RFID trials the analyst firm found around the planet, one out of three are in the U.S. The top 10 countries were the United States (with 456 trials); the United Kingdom (145 trials); Japan (104); Germany (64); China (54); France (45); The Netherlands (36); Korea (32); Italy (26); and Canada (25).

Those numbers can be misleading, though, as they do not necessarily reflect industry investment. In Korea, for example, the government directly provided the money for 15 of those trials, Das said, adding that a government's interest is in strengthening the future economy and less of the quarter-to-quarter, quick return-on-investment mentality that often dominates corporate America and the U.K.

The initial uses of RFID in deployment are not the EPC product identification types that get so much publicity but instead are more practical, such as Mobil SpeedPass and E-ZPass in the United States.

"In the value of the tags and infrastructure, the RFID business remains focused on smart card/payments and key fobs," the report said. "The largest projects have not involved EPC tags, and they include the $6 billion China national ID card scheme and the delivery last year of 50 million RFID tickets for the Osaka World Fair by Toppan Printing using Hitachi Mu Solutions RFID inlets."

The study also challenges another RFID assumption. A high-tech industry that is used to projecting prices according to Moore's Law—Intel's Gordon Moore's theory that transistor capabilities would double every couple of years—assumed that RFID chip prices would fall as deployments kicked in, with the only questions being, "How far and how fast?"

But as companies think through RFID strategies, they are coming up with new uses, often requiring more sophisticated and costly chips. Unlike a PC or a server, the amount of space for even an extra dot of silicon is quite restricted on the typical RFID chip.

The report offers as an example a United Nations effort to make passports more secure and consistent. "The passport scheme involves about 30 countries and 300 passports working at high frequency. It has gotten off to a slow start, but it will rise in importance in the next three years, perhaps reaching 20 million yearly. That may not seem much, but—like most RFID smart cards—these labels and inserts cost 15 times as much as a retailer's EPC tag," the report said.

"The chip alone costs two dollars in a passport because of the large memory, sophisticated security and use of a microprocessor. RFID smart cards also have sophisticated security and microprocessors," the report said.

"The China ID card is an exception as it has hardwired logic and no microprocessor, but it still costs $1.25, 6.25 times the cost of today's retailer's EPC tag, and the gap will widen as EPC tags sell in larger numbers and their costs tumble further."

Another report example is one where the chips need to be more expensive because they need to be ruggedized, so as to withstand extreme temperature, humidity or extensive wear-and-tear. Laws mandating animal tagging for slaughter were given as an example.

"It may be that the high unit price of animal tags will, together with legislation in many countries mandating their use on livestock, lead to this being the second-largest market for RFID tags by value this year," the report said.

"But although we note suppliers of these low-frequency tags with necessarily sophisticated encapsulation for ruggedness, biocompatibility, etc. are doubling and trebling output, that may or may not be enough for demand, and case studies of the tags in action are difficult to obtain."