From wearable IDs to facial recognition using thermal imaging, the military is using state-of-the-art biometrics technology.
The Sept. 11, 2001, terrorist attacks accelerated the growth of U.S. military applications using biometric technologies to identify the enemy and to improve security and surveillance. The adoption of biometrics in recent months has been driven by advances in a variety of technologies, including wireless, artificial intelligence (AI), and machine learning.
Biometrics covers a broad range of identification technologies, including fingerprints, facial scans, and iris patterns, which now offer higher reliability and accuracy. Here are a few advances in biometric technologies led by the military.
The soldier’s wearable authentication token project
The U.S. Army’s wearable authentication tokens have been in continuous iterative development since the original prototype system was completed in 2019. The wearable identity tokens combine public-key–based credentials with advances in the commercial wireless payment industry and flexible hybrid electronics, according to Ogedi Okwudishu, a computer scientist within the U.S. Army Combat Capabilities Development Command in the C5ISR Center’s Tactical Network Protection Branch.
“Using smart cards is not operationally feasible in tactical IT environments,” said Okwudishu.
Like the common access card (CAC) containing biometric data, the tokens will use public-key–based authenticators. Soldiers will use them to connect wirelessly and securely via Bluetooth Low Energy to an end-user device. They must use a PIN as a second authentication factor to get the network system to grant them access to a restricted area. Once out of range of the system, they are logged out.
The tokens are currently at the prototype stage. “Government development will continue through 2021,” said Okwudishu. “Commercial development is planned to start in late 2021, and fielding [field tests] should begin in early 2023.”
The project team is “in the process of implementing and testing in the Soldier Authentication testbed,” he said. For the next several months, the team will focus on [biometric] authentication to Windows and Linux devices, a mobile version of the test token for human factors testing, and the use of “other wireless protocols and technologies under development by the Army.”
Identifying enemies in the dark
The U.S. Army Research Laboratory (ARL) has conducted experimental tests that combine facial-recognition technology with thermal imagery to allow soldiers to better identify persons of interest in the dark. The new technology uses AI, machine learning, and infrared cameras to identify facial patterns by detecting radiated heat from the skin.
In 2019, an initial handheld binocular prototype was developed for short-distance scanning. A more advanced prototype that is capable of imaging at a long distance is being developed, Hu said.
ARL is also working to improve its algorithms. The Army’s corporate research laboratory and a team of scientists have already developed the initial algorithm that ties the data into the integrated software and hardware platforms.
The ability of thermal imaging is made possible by detecting radiated heat from the skin or objects, and when properly calibrated, imaging can measure the absolute temperature from the skin, Hu noted.
The biggest advantage is that thermal imaging can capture distinctive features behind a face mask covering. It allows night-time facial recognition in low-lighting areas. However, the challenge is that the imaging works only with thin, very tight-fitting facial coverings.
Imaging through thick coverings requires technologies such as millimeter-wave imaging like scanners employed at airports; however, such technologies are not in small form factors, said Hu.
Biometrics database modernization
The U. S. Army has modernized a 20-year-old biometric database with a new software update to help soldiers patrolling at foreign checkpoints to identify persons of interest in real time. “The original database relied on software to create relationships between tables,” said William Daddario, an engineer in the CCDC C5ISR Center’s Cryptographic Modernization Branch. “We built the relationships into the database. This is a more efficient way to operate.”
So if one change is made in one area, it propagates throughout the whole database.
The handheld Biometrics Automated Toolset-Army (BAT-A) is used by soldiers to collect and process biometric identification data such as irises, fingerprints, and facial images. The subject can wear glasses or contact lenses while an iris image is taken. The collected data is compared with the data stored in the Department of Defense’s database of 1 million biometric entries.
“The center has performed vulnerability assessments on previous systems for PM DoD Biometrics [the project management office for the Department of Defense Biometrics],” said Daddario.