I got an e-mail recently from a reader in the Netherlands asking for ways to deal with “false-positive” reads—that is, when an interrogator picks up a tag it wasn’t meant to read. “This is often mentioned as a problem in literature,” he writes. “However, I cannot find any information about ways to reduce this risk.”
My correspondent is right. It is common to read tags on products on adjacent conveyors or through adjacent dock doors. One way to deal with this problem is to build a Faraday cage or Faraday shield around the reader antennas. A Faraday cage is an enclosure formed by conducting material, usually metal mesh. It’s named after Michael Faraday, a physicist who built one in 1836 and explained why energy doesn’t pass through the metal mesh. The electrical charges within the conductor repel one another and as a result, the energy travels through the edge, or skin, of the cage.
As a quick aside, I saw an unforgettable demonstration of this in a science museum once (I can’t, for the life of me, remember which one). A man got inside a large metal cage, which was then zapped with thousands of volts of electricity. The audience gasped when he placed his hand on the inside of the cage as the electricity coursed through the outer edge of the metal he was touching.
Now back to false reads. Faraday cages work. I’ve been in more than a few warehouses where companies have built them around interrogator antennas on conveyors to prevent the interrogator on conveyor No. 1 from reading a tag on a case moving on conveyor No. 2.
It’s worth pointing out that false positives are not always a problem. It depends on your application. Wal-Mart doesn’t care if a product entered its store from one dock door or another. It only wants to know the product arrived. So false positives are not an issue.
For applications where false positives are a problem, early adopters are moving away from Faraday cages and other shielding solutions and are instead using software to interpret reads.
Let’s say a tag is read for the first time by an antenna at dock door No. 1, and that same antenna reads the tag 30 more times as it moves through the portal. Another antenna at dock door No. 2 reads the tag after it’s read at dock door No. 1 and picks it up another 6 times. Software can be used to infer that the tag was on a product coming through dock door No. 1.
Similarly, if a forklift picks up a pallet and reads cases on that pallet continually for two minutes, but also picks up other tags that are read only a handful of times, software can infer that the cases read continuously were on the truck and the ones read a few times were not.
This kind of logic isn’t hard to build into software, but software companies need a lot of real-world experience to be able to develop the logic that will work well in different situations. As middleware vendors work with their customers, they will continually improve the ability of their products to interpret tag reads. For instance, software might soon be able to determine if a forklift is picking up a pallet or putting it away. That will make life a lot easier for companies trying to leverage RFID data to improve their operations.