Next-Generation Biometrics
Written by Harrison Donnelly
ENHANCED DATABASE DRAMATICALLY INCREASES
CAPACITY AND EFFECTIVENESS OF IDENTIFICATION
PROGRAMS IN FIELD OPERATIONS.
An enhanced database of biometric information has been operating effectively since early this year and has already led to the identification of an individual associated with IED use in Iraq, Pentagon officials said recently.
Members of the Biometrics Task Force and other Department of Defense officials in February announced replacement of the current Automated Biometric Identification System with an enterprise-level, multimodal database, the Next Generation Automated Biometric Identification System (NG-ABIS), capable of storing greater amounts of data and supporting new forms of analysis.
While biometrics is increasingly used in a wide range of DoD functions, the new system will be especially important in supporting operations in Southwest Asia, where the technology has been extensively deployed to control base access and target suspected insurgents.
“Biometrics removes the terrorist’s secret veil,” said Army Colonel Theodore L. Jennings Jr., project manager, DoD biometrics, within the Army Program Executive Office Enterprise Information Systems (PEO EIS). “It gives the warfighter the ability to tell friend from foe in the field.”
The Biometrics Task Force and PEO EIS oversaw development of the new database, which was carried out by an industry team led by Northrop Grumman using existing technology. “This is a great example of how to adapt commercial technology and get it into the field quickly. This is an example of how to do it right,” said Deputy PEO EIS Lee Harvey.
NG-ABIS will be the central, authoritative, multimodal biometric data repository. The system operates and enhances associated search and retrieval services and interfaces with existing DoD and interagency biometrics systems. The repository interfaces with collection systems, intelligence systems and other deployed biometric repositories across the federal government.
NG-ABIS’s greater functionality will expand the ways the data can be used and will serve as the strategic level authoritative data source for DoD biometrics. New software algorithms and additional modalities will provide improved response times and better matching results.
MULTIMODAL MATCHING
The scalability of the system is critical as more people are included and as the list of biometric methodologies expands beyond the current base of fingerprinting, iris recognition and facial scanning. That ability of the NG-ABIS to expand is provided by IBM BladeCenter technology, which offers high-capacity processing along with reduced requirements for system rack space and data center footprint.
The NG-ABIS also offers faster operations, providing customer response times that are from 14 to 28 times faster than the previous prototype. In addition, electronic transaction times decreased by a magnitude or more, while numbers of “yellow resolves”—uncertain outcomes requiring human intervention— decreased by more than 50 percent.
Developers also emphasized the value of the new system’s greater processing capacity and improved algorithms in finding matches between biometric information coming in from the field and existing records of individuals. In the case of the Iraqi IED-maker, for example, the identification was based on 11 partial fingerprints found on the debris of exploded bombs, which when combined through a formula targeted a specific person with a high degree of certainty.
Even more significantly in the long run, the new capacity holds out the promise of multimodal matching, which can determine an identification based on results from different forms of biometric technologies.
That ability is important because any one technology can produce three different types of matches of an individual to the database: positive—the person is definitely on the list of individuals to be apprehended or excluded from a facility; negative—the person is definitely not on the watch list; and inconclusive, requiring further investigation.
Thus, a fingerprint, iris scan and palm measurement of a person seeking entry into a U.S. base in the current area of operations might each produce a partial or uncertain match with a person on a list of known terrorists. But use of a multi-modal fusion algorithm could provide a definitive decision based on the several different uncertain identifications.
Officials said they were not aware of any other implementations of multimodal matching currently in the world of biometrics. ♦
