No silver bullets

Hype initially surrounded Defense Department plans to use passive radio frequency identification (RFID) technology to manage and track shipments from 60,000 suppliers in DOD's global distribution chain. But reality has tempered the excitement during the past six months.

Air Force, Army, Navy and Defense Logistics Agency officials have tested RFID technology, and demonstrated that it cannot magically solve supply-chain problems. In some cases, the use of RFID technology creates a new set of problems that will require changes in basic DOD business processes, department and industry project managers say.

On the regulatory front, DOD has been hobbled by slow development of regulations that will mandate suppliers' use of RFID technology. After an 11-month delay, Pentagon officials issued proposed rules last month that name the first types of supplies that must use RFID tags.

At first glance, passive RFID appears to be a silver bullet that can help DOD meet its goal of efficiently and accurately tracking billons of items that pass through its supply chain.

DOD employees identify items in the department's distribution chain with bar code labels. But bar code readers, which use beams of light to read a label, only work if the beam can reach the label. That means boxes or cases stacked on a pallet need to be unloaded and then scanned individually.

Passive RFID theoretically solves this problem by storing information on a 64- or 96-bit computer chip embedded in a substrate inside a paper tag sandwich that usually sports a printed bar code, too. A reader, either fixed or handheld, beams a low-power signal in the UHF frequency band (860 MHz to 960 MHz) to access information from tags on pallets and the individual cases stacked on those pallets at a range of about 3 feet.

In real-world tests, however, that theory did not meet the predictions. Larry Loiacono, an information technology specialist at DLA's Defense Distribution Center in New Cumberland, Pa., said tests this year at the two main distribution depots in Susquehanna, Pa., and San Joaquin, Calif., showed RFID was not ready for prime time.

Stationary or "portal" tag readers from Symbol Technologies mounted on four warehouse doors in Susquehanna and five doors in San Joaquin were able to read the pallet tags, also provided by Symbol, almost 100 percent of the time. But that rate dropped to unacceptable levels, Loiacono said, when the readers scanned tags on cases stacked on the pallet.

The read rate on cases was about 80 percent, Loiacono said, which would not support an electronic data interchange transaction that would generate payment to a supplier, another important part of the grand DOD RFID scheme. "I'm not prepared to send a payment to a vendor" based on such a low read rate, he said.

These sentiments were echoed by Dave Cass, director of the Ocean Terminal division of the Fleet and Industrial Supply Center in Norfolk, Va., operated by the Naval Supply Systems Command. During a presentation at the Air Force Supply Chain Automatic Identification Technology Forum held in St. Louis in March, Cass said DOD users cannot expect 100 percent read rates on all shipments all the time.

"Business processes must accept this limitation of the technology upfront," he said.

RFID is not acceptable for transactions of record, Cass said during his presentation. "Human interaction is required to provide a check and balance that everything that should have been read was read."

Despite repeated requests, command officials did not make Cass available for an interview. They provided a report, however, on the Ocean Terminal's RFID system. The report states that although Ocean Terminal officials had planned to use an unmanned, stationary RFID portal, they quickly decided to use a manned portal to improve accuracy of read rates. In his St. Louis presentation, Cass said he also deployed bar code scanners to improve tracking accuracy.

Air Force officials experienced extremely low read rates on RFID-tagged military shipping labels at Dover Air Force Base, Del., and Ramstein Air Force Base, Germany, in tests conducted by Northrop Grumman that ended last December.

Michael Moore, defense enterprise solutions manager at Northrop Grumman's AIT Center in Williamsburg, Va., speaking at the Air Force conference held in March in St. Louis, said that the read rate of Symbol tags before consolidation at Dover was 83 percent for 676 tags. At Ramstein, before the pallets were broken down, the case read rate dropped to 65 percent, he said during his presentation.

Peter Langworthy, senior program manger at the Northrop Grumman AIT center, said those low read rates indicate that DOD needs to do more than throw RFID technology into the supply chain. Instead, DOD officials must examine each step of the business process to determine the best way to use tags and readers.

This could include simple fixes, Langworthy said, such as instructing a forklift operator to drive slowly through a stationary portal or adding backup portals for second or third reads to improve accuracy. Langworthy said the fastest way to resolve the poor reads of cases stacked on a pallet was to read the information from tags on each case into a master tag on the pallet, which would improve overall accuracy because pallet tags have close to 100 percent read rates.

For certain commodities, DOD users need to experiment to determine optimal placement of tags on a case to ensure the best read, said Stephen Moody, Army chief warrant officer and team leader for the advanced process and packing team of the Army Soldier Systems Center's Combat Feeding Directorate in Natick, Mass.

Directorate employees conducted a test of RFID to track cases of soldier meals in February 2004. Moody said he discovered that the meals were RFID-unfriendly because of the foil packaging and the food's water content. Both interfered with the low-power RFID signal.

But directorate officials discovered that if there was an air gap at the corners of the end of a case, Moody said, read rates dramatically improved. The test demonstrated that tags cannot be slapped randomly on cases. Instead, they need to be placed in optimal spots determined by trial-and-error experimentation.

Salil Pradhan, chief technologist for RFID at Hewlett-Packard's HP Labs, said shipments with a high-metal content, such as the company's computers and printers, are passive RFID-unfriendly and need well-placed tags to improve read rates. Pradhan said a piece of Styrofoam between objects with a high metal content and the RFID tag creates an air gap and improves the read rate.

Alan Melling, senior director for electronic products code solutions at Symbol, said even in the best of circumstances, read rates will never be 100 percent. To reach the optimum read rates, officials must consider revising their business processes and perhaps buying additional portals, readers and antennas, he said.

For example, Cass improved accuracy at the Ocean Terminal by supplementing the RFID tags with bar code scanners.

But, Melling added, any enterprise that deploys RFID also must realize "there has to be a manual exception process," in which someone enters information if nothing else works.

Besides changing business processes to improve RFID rates, DOD and its suppliers also have to overcome regulatory hurdles. DOD officials were supposed to issue proposed rules last May mandating the use of RFID on a narrow class of products, including packaged meals, clothing, equipment, tools, tents and repair parts.

But DOD officials did not publish the proposed Defense Federal Acquisition Regulations Supplement covering those supplies until late last month. Suppliers and industry have until June 30 to comment on the proposed regulations.

The spokeswoman for Alan Estevez, assistant deputy undersecretary of defense for supply-chain integration, did not return calls asking for a time estimate on a final rule.

But Megan Games, manager for defense opportunities at Input, a market research firm, said Estevez and his staff could need six months to go through the comments and make any needed adjustments before issuing the final rule in October, a one-year delay compared with the original deadline.

Estevez will then need to scramble to develop new regulations for using RFID tags on a wide range of materiel, including ammunition, pharmaceuticals, packaged petroleum, chemicals and construction and barrier supplies. Their goal is to use RFID tags by next January on those supplies, which are delivered to more than 30 Army, Navy, Marine and Air Force depots.

Yet another regulation will have to be developed by January 2007 for all other classes of supplies shipped to DOD. Department officials originally planned to use RFID technology to tag and track all supplies from 60,000 vendors by then.

Despite the technology and regulatory obstacles, Moody remains an optimist. As one of the early pioneers of RFID in DOD, Moody said the RFID technology will help transform the DOD logistics system. "The timing could not be better because the whole Army is in a process of transition," he said.