RFID - Radio Frequency Identification

Radio frequency identification (RFID) asset tracking has made its way into art handling. Fine Art Shipping began working with it in early 2009, and we are in a fine-tuning stage with the system. RFID shares barcoding’s facilitation of electronic databases in inventory maintenance, but takes the concept a step further. The idea is to improve the accuracy and efficiency of our storage maintenance operations by eliminating several points of potential human error and object handling. And apart from the nearly completed updating of our existing storage with RFID, it has already provided savings on some jobs by streamlining every stage in tracking our custody of objects - up to and including billing.



RFID technology has been industrialized since 1970 with large-scale tracking operations in such industries as mega-farms and railroads. In recent years it's been used most visibly in the security systems of big box stores. It's now seeing all kinds of applications, from DIY projects in the home to tiny human implants. Most notably for our purposes, it's a cutting-edge alternative to barcode-only asset tracking systems. For those who are tracking the movements of objects by hand and checking off hard copy lists, RFID represents a great leap forward.

The beauty of RFID is that the transponder tags don't need to be visible, or even accessible, once they are placed on or in an object's packaging. Our readers can pick up any number of RFID tags while they are packed inside a crate. They can also potentially read through multiple packages to scan an entire shelf of storage. (Results will probably vary around large objects composed of materials that block radio waves, such as large steel plates.) By the time you read this, FAS will have begun testing the more precise limits of this feature in our equipment in order to determine the minimal tag placement conditions necessary for scanning coverage.


How does non-line-of-sight scanning not crowd the readout with unwanted items? Simply tell it what to look for. Entering the job number(s) associated with the tags that you are seeking, or the container numbers, or the range of inventory numbers, narrows the reader's parameters down to the wanted tag or set of tags. Once the tags are found, we can add more details or status updates to their associated files in the storage database as we physically handle the packages. The relocation of an inventory from one storage rack to another can be typed into the reader by hand while it's being moved, or the racks themselves can be scanned as the new location. (Our storage racks are tagged for location data.)


The equipment that you actually handle consists of two parts: a reader and a tag (or 100,000 tags). Our handheld reader is a mobile computer. We can take it into the field and record the movements of labeled objects as they are loaded into our trucks or delivered to consignees. This can be performed for multiple jobs between syncing, but like a smart phone, the current model of handheld reader must be returned to its dock in the office to upload the new status of the objects and sync with the central system. Another option is taking one of our networked laptops along. Thanks to the fact that our database utilizes cloud computing, we can look up or enter information from any online computer or mobile device. The next software update of our handheld reader will catch up to this potential and write to the cloud database in real time without returning to a sync dock.



The RFID tags that we use are sticker labels. The transponder is a thin circuit on the adhesive side of the label, composed of a microchip and antenna, and the face of the label displays any visual information wanted; including bar codes, job and client information and item numbers specific to the job. For example; every job can have a numbered list of objects starting with 1, 2, 3 and so on, with those numbers printed on the labels. But the RFID assigned to each label will remain unique. This unique number is also printed on the face to support other details in visual identification. Adhesive-backed plastic pouches make these labels easily removable from an object's packaging. They also make the labels reusable, since RFID tags can be reassigned to new entries in the storage database at any point.


A single artwork can be made to stand out from a large inventory by temporarily assigning it a new job number from within the database. Then the reader will only pick up that object when its temporary job number is entered into its parameters. Or the object can just be selected from the related job's list of associated items on the tool's full-color touch screen. This is a tremendous time saver when a specific object needs to be pulled from long-term storage. We can quickly consult the database for its present location, as well as the size and shape of the package being sought. If the package is not initially visible at the physical location, the reader can be used to quickly confirm its presence there and narrow down its exact position to a few cubic feet. A search that might have formally taken a day or two is now completed in minutes.


We have a more powerful type of reader on a podium that floats around our warehouse docks. This one generates a larger electromagnetic field (a portal) through which no tagged object can pass without its arrival or departure being recorded. A storage account begins the moment the inventory arrives through the portal and ends when it exits, and the exact duration between those events is immediately made available for billing purposes. Multiple mobile laptop workstations are also networked to the database, allowing us to add detailed information (such as dimensions, artist, title, etc.) to the object being scanned and guide how the status of that object is being changed. Photographs taken from networked digital cameras are automatically sent into the correct object entries in the cloud database.

The camera is admittedly a bit slow when dialed into the database, and sometimes throws up an error message. While Erich works on this issue, he and Ron are taking the opportunity to research wireless options, so that the camera will no longer need to be physically wired to a computer running the storage maintenance system. We can still take photos separately and disperse them into the database by hand, but getting the connected camera up to speed will send images automatically into the correct object entries and eliminate that aspect of database management. [Update 5/27/10: Erich has modified the software, and our cameras are now connected wirelessly to the database, automatically adding photos to the open database entry.]




We use a passive RFID system, which can also be described as a system using passive tags. This means that the RFID tags have no internal power source. Their unique radio signals are triggered only when they fall within the field of a searching reader. This requires a higher level of power in the reader than would be needed in an active system, and limits the maximum range between reader and tags to 3-5 meters at best; about 16.4 feet max. It also limits the number of readings within a few meters to a factor of hundreds.

An active or semi-passive system would use battery-powered tags, reducing the power needed for the reader and greatly increasing the effective reach of the scanner's field to hundreds of feet and the quantity of readings to thousands. Active tags also emit radio frequencies constantly. One major difference between the two types of systems is that active tags transmit their own signals on a constant basis and may include different kinds of sensors. This makes them suitable for use when a higher level of RFID-based security is desired in addition to an auditing function only. One company ISIS states that their products provide "...an early warning system that gives security controllers the ability to prevent theft rather than merely record it." This is possible because individual tags with vibration sensors send a real time signal as soon as the object it is attached to is touched. An examination of this type of RFID will require a separate article.

Passive tags are relatively inexpensive (about one fortieth the cost of active tags), and don't need to be physically maintained with battery replacements. I haven't found any information on how long our passive tags are expected to last, but the consensus appears to be "a long time", whatever that means.

At this time, the economy of our system is proving well worth the lower area of affect in the readers for our purposes. For one thing, we typically use our readers to look for specific sets of tags. And even if we need to perform a complete inventory of every object in our warehouses, it doesn't need to be completed in one quick wave of the reader. We are close to having our entire storage refitted with RFID tags, and once that is done, I plan to test how long it takes to make a complete inventory update. This would take at least two people several days (if not a few weeks) with pen and paper checklists. Not to mention the potential risks of opening packages and handling objects. I'm predicting that we'll be able to do it with one person in a few hours, and without a single package opened.

There are read-write tags, to which you can add or overwrite ID information; WORM tags (write-once, read-many - meaning that they can be overwritten once) and read-only tags. Our current tags are technically read-write, but we are using them more-or-less as if they were read-only. Our software gives us so much flexibility in adding to or overwriting their associated files in the storage database that we have no need to change the unique signatures of the tags themselves once printed. The reason we use read-write tags is that they allow us to assign our own numbers to them when they are printed (and coded), rather than using a generic pre-coded numbering sequence from the manufacturer. We combine the job number with that job's item number for a unique but immediately identifiable ID on each object.

So here's an example of another way that we use the system. Say we need to pull 83 objects from storage to be installed at a site. We assign that portion of our storage with the job number for the task. We enter that job number into the portal so that it knows what to look for, and a list of the 83 objects appears on the laptop connected to the portal scanner. We load our truck through the portal, and the line for each object is highlighted in yellow as it passes through - indicating that it is leaving storage. On site, the same job number is entered into the handheld scanner, pulling the same list of objects up, now highlighted in yellow. As the truck is unloaded on site, the handheld scanner is swiped over the objects changing each object's line on the screen from yellow to green. When every line item in the list is green, the objects are all accounted for as delivered.

Lately I've been thinking about the potential to develop our own software for transforming smart phones into RFID label readers via barcodes. The barcode reading apps already available for smart phones may prove economical when the handheld scanner is in one location, and we want to scan objects in other locations at the same time, using line-of-sight. Better yet, here's a recent DIY article on turning your smart phone into a bonafide RFID scanner. Smart phone scanner/readers would take advantage of cloud computing's scalability, which I will address in the next article.


Incidentally, I've read one opinion that the tech behind contemporary RFID systems can be traced back to 1945, when Léon Theremin [Lev Sergeyevich Termen] abruptly returned to the Soviet Union from his home in New York City under unknown circumstances to develop the infamous surveillance device nicknamed the "bug". But why not to 1920, when he invented his eponymous musical instrument? The theremin is played without physical contact by interrupting two electromagnetic fields with your hands - no transponder tags needed.


In 2009 Timo Arnall and Jack Schulze began mapping the invisible fields of RFID readers on film in an information design collaboration between Touch and BERG. Immaterials is an interesting video, and I'd like to see larger versions of these experiments with the more powerful scanners that surround us.

Next: SMI - the software that makes RFID an art handling technology.