What's New in Machine-Readable Technologies

General
Typography
  • Smaller Small Medium Big Bigger
  • Default Helvetica Segoe Georgia Times

While the technologies enabling machine-readable data are momentarily stable, ideas for capitalizing on existing technologies are exploding.

 

"Machine-readable" technology, by definition, has been around since Thomas Edison gave us the phonograph record. In recent years, we've come to think of this field as being dominated by optical-character recognition (OCR) for scanning standard documents, magnetic-ink character recognition (MICR) for checks and financial papers, and radio-frequency identification (RFID) for tracking the locations of goods and objects. However, the whole field encompasses the concept of automatic identification (AID), the process of helping machines identify objects, capturing data about those objects, and making that information available as data without requiring manual data-entry procedures. This definition can include such objects as smart cards and such technologies as biometrics. It also includes ways of letting common terms be "synchronized" in their meaning throughout related data sets.

 

Although we're regularly bombarded by news of this or that innovation in other technological areas, AID technology is currently in a phase where the technology itself is, for the most part, stable. Where the envelope is being pushed is in the devising of new applications for these existing technologies, and the ethical and legal implications of those choices.

RFID: Tracking Merchandise and Materials

Although it's been around since the 1970s, RFID is the part of this field that is currently most prolific when it comes to new ideas for applications. Long a staple of enterprises, particularly in the manufacturing-to-retail supply chain, for identifying locations of goods and their movements, RFID tags come in three major versions. The most common is the passive RFID tag, which draws power from the radio field generated by an external reader to activate the tag for reading and has a range of about one foot. Newer is the active RFID tag, which has a small internal battery that lets readers pick up the signal of their information from farther away (three to 20 feet) and allows reading of multiple tags at once. The third version is the hybrid battery-assisted passive tag, which also has a small battery, but one that is activated only when it senses a tag reader is nearby.

 

The radio waves needed to make the technology work can pass through non-metallic materials, and the microchips in the tags can contain unique serial numbers. This makes the chips ideal for attachment to, for example, consumer goods, and tracking them until they leave a store in the customer's hands.

 

Despite its longevity as a technology, RFID adoption has been limited until relatively recently because of several stumbling blocks. The main one is expense. For example, active tags that operate at ultra-high frequencies (UHFs) can be read up to 300 feet away, but are relatively costly and therefore are usually used for tracking entire lots of goods rather than individual items. A second problem is lack of international standards in certain RF aspects. Now that supply chains are often global in scope, the fact that the U.S. uses 915 MHz for UHF, Europe uses 868 MHz, and Japan wants to use 960 MHz means there's no international standard on frequency use. On top of that, there are other electronic devices that use those wavelengths, with which UHF tags and readers could interfere.

 

Accompanying the spread of RFID is the use of software agents that run on the computers to which RFID readers report their data. These agents can handle input from multiple readings faster than humans can and can make simple decisions about routing information and whether or not a human needs to take a look at a particular situation.

IDing More Than Just Your New TV

However, RFID is growing beyond the traditional supply chain environment, and the implications are bringing a growing tinge of controversy to what might ordinarily seem a pedestrian industrial procedure.

 

We've all heard heartwarming stories about the recovery of someone's dog or cat because of an RFID chip planted in the animal. Less well-known are parallel uses for livestock. Similarly, some school districts have started embedding chips in the clothing or ID badges of children in order to track their location and movements.

 

The U.S. Food and Drug Administration (FDA) has approved implanting of RFID chips in people with certain ailments (e.g., diabetes, Alzheimer's, cardiovascular disease), the idea being that if those people are taken to a hospital unconscious, the chip could contain a reference number to a central database through which the patient's medical history might be quickly obtained, hastening critical treatment.

 

A related use of microchips is in machine-readable travel documents (MRTDs). Functioning like a smart card, epassports (in use in a number of European nations and limited use in the U.S.) have embedded in them microchips with between 32 KB and 512 KB of memory and an antenna, which lets the document store at least the basic information from the passport, and at the higher memory level, biometric information (e.g., fingerprint scans, retinal scans) about the passport user, which can be read by appropriate equipment. Several European countries are also using microchips in personal identity cards for their citizens and residents.

 

While it's clear that applications like this are potentially lifesaving (or at least useful), it raises issues of privacy. Could the government, or businesses from which we buy goods, track our movements or locations? EPCglobal, an international standards organization, maintains an Electronic Product Code Tag Data Standard, which offers commands to permanently disable RFID tags on purchased goods, but while many retailers do this voluntarily, not all do. This danger can be avoided by consumers themselves removing or destroying RFID tags from newly purchased items or discarding packaging that contains the tags. However, this may not be an option in the future if tags are embedded in clothing or other goods as part of the manufacturing processfor instance, something that is already being done by Sweden for its army uniforms and is planned for other military equipment.

 

Today there are existing problems with theft of information from smart cards, which use similar technology. A quick scan through the gadget catalogs that flooded our mailboxes just before the recent gifting holidays shows a profusion of "secure" wallets, badge holders, and other means of protecting smart cards from unauthorized reading. MIFARE, a product of Austrian company NXP Semiconductors offers "Generation 2" UHF tags, which combat this problem by encrypting data transmitted from card to reader, although so far it's a proprietary technology that hasn't been widely adopted and has a reading range measured in inches.

The Smartphone Connection

On the other hand, it's getting to be more practical to track people via their smartphones, which is easing concerns that it could be done via RFID specifically. But it seems clear that privacy issues will be part of a cultural adjustment in coming years because of smartphones, RFID, Web activity tracking, and other innovations.

 

As it happens, though, smart phones may also be the next big thing for RFID use as well.

 

Thinaire, a cloud-based software platform that reads the company's Smartrac RFID tags, was recently released by a company of the same name as the product. Smartrac tags can be attached to manufactured goods and accessed by a consumer's smartphone that's equipped with Near Field Communications (NFC) abilities, through which the phone user can access additional product information. This can include information about style and whether a specific color or size is available at that store or other stores nearby. Alternatively, it can steer the potential buyer to the manufacturer's online site for a wider selection.

 

NFC capabilities are also in use at a Boston-based advertising agency, Allen & Gerritsen, to indirectly encourage marketing strategies. The Pic Tap Toe project uses photos posted via Instagram, in conjunction with RFID technology, to let employees at a newly acquired Philadelphia ad agency get to know employees at the home office in a digital sense. Employees in both cities post photographs of their home cities and personal lives, then use X and O pieces with embedded RFID tags to play Tic Tac Toe between offices on wall displays. While defending the game's apparent frivolity as a way for the two staffs to get to know each other, A&G plans to use the technology in the future to, for example, create loyalty cards that activate screens showing discounts and special offers when a consumer enters a participating store.

 

Other RFID use ideas announced in recent weeks are Metalcraft's RFID Wristbands that are designed for crowd and access control at events, SensMaster's Tristan wristband tag that checks personnel safety conditions, and Versus Technology's Advantages Clinic Real Time Location System, which is a wristband that guides medical patients to vacant color-coded exam rooms and tracks how long the patients have been waiting to see medical personnel.

Big Data and Machine Readability

Aside from the AID aspects of machine-readable data is the problem of drawing usable information from databases and data sets that are using different structures and running on different computer platforms. While this has always been a challenge in enterprises using multiple platforms, it will become much more prominent in the future as we move into the era of Big Data.

 

Most standard database and spreadsheet programs use the Comma Separated Variables (CSV) format, which is text-based, to facilitate transmission between machines but doesn't include metadata information. This can cause problems if the same data is referred to in different ways across data sets. A simple example would be if the United States is referred to as something like U.S., U.S.A., or America in different data sets. Software looking to correlate information may not include logic capable of determining that those terms can refer to the same nation. (Yes, it should, but this is a simple example.)

 

That's where the Resource Description Framework (RDF) standard comes in. Invented by the World Wide Web Consortium (W3C), the RDF provides a means of conceptual description or modeling of metadata without regard to specific syntax or data formats. Related to eXtensible Markup Language (XML), it's been around since 1999, and it describes metadata in terms of graphs rather than the more usual tree structure. But its capabilities to attach metadata to data sets and provide a means of relating common terms to each other across data sets are becoming increasingly important as larger data groups are analyzed to compile useful information.

 

Another advantage the RDF provides in Big Data applications is Web syndication, a means of making material from one Web site available to multiple other sites. The most common formats enabling syndication are Atom, an XML language for Web feeds; JavaScript Object Notation (JSON), which is related to the C family of languages and can describe name/value pairs and ordered lists of values; and Rich Site Summary (RSS), which syndicates data between Web sites automatically. As the Internet becomes more interlaced, such techniques as the RDF can only grow in use.

 

When dealing with information offered by the U.S. government, there is the open-format standard described in the Open Government Directive of 2009, which requires that "agencies should publish information online in an open format that can be retrieved, downloaded, indexed, and searched by commonly used web search applications. An open format is one that is platform independent, machine readable, and made available to the public without restrictions that would impede the re-use of that information." This standard is compatible with Web syndication schemas like the RDF and will figure in Big Data applications that use U.S. government data.

Biometrics for Greater Computer Security

A final area that the machine-readable technology umbrella covers is biometrics. Biometrics refers to various means of identifying individual humans as authorized users of particular computer resources, other than the standard methods of passwords and personal ID numbers. While we've touched on its use in epassports and similar identity documents, it has a fuller future in authorization and authentication of computer users.

 

Biometric technology interfaces directly to computers and, depending on the type of reader, can identify users based on fingerprints, palm prints, facial recognition, retina or iris scans, and keystroking rhythms. The expense of readers and interpreting software, as well as privacy concerns about sharing biometric identifying information because it's unique for every human, has suppressed the spread of this technology.

 

However, as the technological indoctrination of the human species grows and it becomes harder for human minds to remember all the passwords necessary to access our tools and toys, biometrics will have to play a growing role in the future. Identity theft is mushrooming, and biometrics is an effective answer to controlling access to technology resources. Eventually the economic payoff of prevention will trump the resistance to its startup expenses.

 

The new frontier in biometrics is addition of biometric features to mobile devices, in addition to standard computers. Goode Intelligence, a market-research company, predicts in its "Mobile Biometric Security Forecasts 2013-2018" report that by 2018, 3.4 billion users will have biometric features on their mobile devices. Frost & Sullivan's Technical Insights research group announced December 5 that it foresees fingerprint recognition remaining the major biometric recognition method for at least the next several years. Also during the first week of December, Apple announced a patent for facial recognition for the iPhone, a fingerprint reader for the iPhone5, and acquisition of 3D sensor manufacturer PrimeSense, which was a key supplier for early Microsoft Xbox Kinect playstations.

 

While the developmental ideas we've reviewed here don't extend the actual technology of machine-readability much, it shows that the technology we already have is barely to the toddler stage of development. Indications are that ideas for its applications could one day carry it close to ubiquity.

BLOG COMMENTS POWERED BY DISQUS