RFID is stupid

We’ll soon live in a country where sensors outnumber people. If you believe what you read, RFID is a magical technology that will forever banish wires, grocery store scanners, credit cards and pocket change from our working and personal lives. There will be no more theft, no more injuries from improper medication, no more lost remote controls, and no more stray dogs. Everything that moves will have an RFID tag either stuck to it or stuffed under its skin.

While Sun Microsystems and others get us all worked up about sensors, equating sensors with RFID, people who work on and research embedded technology are having a laugh at consumers’ expense. Sensors, indeed. Passive RFID, the most common type because it’s the least expensive, costs about 7 cents per tag. But that doesn’t include back-end costs and expensive readers. People who design systems for factory floor automation, logistics tracking, warehouses, external and implantable medical devices, and other blue collar and white coat locales understand that RFID is a limited technology with limited, albeit important, applications.

I have a lot of reasons for being no fan of RFID, but RFID has one serious, show-stopping shortcoming that trumps the others: It’s stupid. A passive RFID tag is incapable of learning, logging, sensing the world around it, or doing anything on its own. If the tag is separated from the reader, or the reader is separated from the back end, the system is going to miss something. Active RFID, which can incorporate sensor-like capabilities, requires a battery or other source of power. When you cross that line, you’re not in RFID-land anymore. You’re just sending and receiving wireless data. The cost rises, and smarter solutions are within reach.

To illustrate, consider the oft-cited use of RFID for hospital patients, to ensure that they get the proper medication and that they don’t get a kidney removed when they’re wheeled into the OR for a blown-out knee. Knowing whom you’re dealing with is dandy, but there’s a lot more you’d like to know about a patient, like heart rate, respiration, temperature, blood oxygen, mobility, and how much noise visitors are making. All of this data can be gathered and recorded by one microcontroller or custom programmable logic device; I’ve found specific technical notes from semiconductor vendors for each of these applications. I like the idea of encrypted and authenticated delivery, too.

In modern commercial packaging, the core of this type of smart monitor, and the lithium battery that drives it (for as long as five years), would be weightless, and I’d estimate the per-patient cost to be about $4 using off-the-shelf parts. That would not replace the room full of equipment that takes very precise measurements, but imagine the cost savings of eliminating just one visit from a nurse per day by using a smart monitor between checks.

A basic programmable ID bracelet could morph into an adhesive patch that’s placed on the chest to monitor heart rate, respiration, and chest sounds. A programmable sensor integrated into a chest tube could alert medical staff when it’s leaking. Yet nobody dreams of taking RFID that far.

If a technology doesn’t inspire you to dream big dreams of things you can’t do today, why invest in it? Dumping dough in RFID only takes money away from the R&D being done to create more capable, intelligent solutions.