Monday, October 15, 2007

RFID for libraries: HF or UHF? (1)

Using RFID tags in libraries has become common nowadays, at least for public libraries. The high circulation rates make the math simple: introducing self-service for patrons increases the efficiency significantly enough to pay back.

For research libraries with a much lower circulation rate, this advantage is not that great, and as a result, RFID adoption is trailing. But, as a famous dutch philosopher would say, every disadvantage has its advantage (and vice versa...).

HF: the de facto standard

There are several types of RFID, each operating in a different wave spectrum. The most important ones are HF and UHF, operating respectively the High and the Ultra High Frequency bands. Though the basic technique is the same, the different wavelengths make quite a difference in the details.

The de facto standard for library RFID use is to use HF-based techniques. This can be considered proven tech these days, and there is a healthy marketplace with numerous vendors offering systems. This is historical: when libraries started adopting RFID, HF was the more mature technology, and it was foreseen that its shortcomings for inventory would be solved in the future. This, however, has not happened. HF RFID works well for patron self-service, but is still not reliable enough for inventory.

For research libraries, the advantage of introducing RFID are much less in the area of self-service, and much more in warehousing. Research libraries keep their books for a much longer time, if not indefinitely, and have large closed stack warehouses. This is where it gets interesting.

UHF vs HF

In recent years, UHF technology has also matured, and has become the RFID type of choice for industries where tracking and tracing is the most important goal. All the big name-projects use UHF: Wal-Mart, Metro, Marks & Spencers', as well as numerous other ones. UHF can scan hundreds of objects per second, and more importantly, it can do it reliably. And here we come to the big difference: the way read failures are handled. With HF, failures are caused by tags too close to each other, parallel, to the shelf or a wall. These tags are effectively skipped.

UHF also has its share errors. However, the tags are much less prone to distortion caused by being placed too close together; and also the reading speed makes the scanner effectively retry reading a difficult tags numerous times. The main reading errors of UHF are caused by field distortion. Certain types of materials and shapes can work as a conduit, effectively extending the field in which the reader operates in a seemingly random direction. When that happens, the reader will also pick up tags from a number of meters away, rather than just the ones close by.

This causes its share of problems - but they are of a different category. The importance is that where HF fails with silent reading errors, UHF fails by reading too many. Reading tags that turn out to be several shelves away are a problem when looking for a specific misplaced book. However, silent errors are deadly when taking inventory. When you have too many results, you can try to filter the unwanted ones out; when you have too few, there's nothing you can do.

UHF has the promise to deliver. Yet HF is the proven technology. What to do?

To be continued...