Printer Friendly

From sci-fi to daily life.

Biometrics, hand geometry, retinal scan . . . the terms sound like medical procedures or bizarre forms of torture, not the latest in access control and security technology. Surely pain is involved. The procedures are painless, however, and the industry is certainly feeling no pain as sales skyrocket. According to Personal Identification News (December 1989), the industry had its greatest growth rate ever in 1989. The number of units shipped rose 167 percent, and dollar sales showed a 90 percent increase, hitting $8,650,000. Unlike past years, last year's major industry events involved the integration of biometrics into access control, law enforcement, and information security application systems instead of the introduction of a new prototype device fresh from research and development. Personal Identification News predicts biometric equipment sales will surpass $20 million in 1990.

The stuff of Star Trek and James Bond is part of daily life for many people in 1990. Where else will it crop up in the years to come as it makes its way from sci-fi novels and television shows to every small town in the United States? One possible biometric application is cruising through testing and development now on its journey to local Division of Motor Vehicles (DMV) offices. Like all changes and attempted improvements, the California Personal Identifier Project started with a problem that needed a solution. Gary Murata is manager of the technology applications section of the California DMV and manager of the personal identifier project. According to Murata, this story began with California DMV Director Del Pierce, who came from a law enforcement background to take over the position in February 1986. Pierce examined the DMV'S existing goals, evaluated whether they were being met, and set some new goals for the organization. During this evaluation, Murata explains, "Pierce wanted to know how many drivers there were in the state, and we told him 20 million. He said, Yeah, there are 20 million records, but how many drivers are there?"'

This exchange raised the question of whether the DMV was accidentally issuing multiple commercial driver's licenses. The DMV had its management analysis shop look at the problem, and Murata himself was pulled into the project in 1986 to examine possible solutions. The DMV documented its own licensing processes as well as biometric processes, and tried to determine if the two were compatible and could be integrated in some way. The state wanted to ensure that each person in California could establish only one identity.

"There's no real documented proof that some individuals can duplicate licenses," Murata points out. The fear, however, is that some drivers do and then spread their citations over licenses in different states to avoid suspension or revocation. Negligent drivers could be hauling dangerous cargo such as radioactive material, for example. A system that allows all 50 states to communicate and compare commercial licensing information would cut down on this danger.

About the same time, the number of motor vehicle accidents involving big rigs and buses increased, and congressional concern sprang from public outcry over the growing number of accidents. Congress passed the Commercial Motor Vehicle Safety Act of 1986, which called for a standardized way of classifying commercial drivers and of reviewing and updating tests for license qualification. It states that each commercial driver can be issued an operator's license in only one state. The act also established the Commercial Drivers License Information System (CDLIS), which serves as an information clearinghouse on commercial vehicle drivers. Presently, CDLIS includes information that can easily be compromised or falsified, such as drivers' names, addresses, and social security numbers. In addition to these measures, the Federal Highway Administration (FHA) made grants available to research ways of eliminating duplicate licensing.

California won a grant to research the feasibility of using biometric technology with the stipulation that it operate through a multistate steering committee hat included representatives from Texas and North Carolina, the other bidders for the grant. Other committee members are from Indiana, Connecticut, Arizona, and Illinois. The FHA wanted all regions of the United States represented, because any legislation resulting from the study will affect all 50 states. The steering committee was put together by the American Association of Motor Vehicle Administrators (AAMVA), which as a national organization was able to gather representatives from all over the country.

The steering committee conducted a feasibility study and decided that biometric technology was a likely answer to the licensing problem. In December 1988, the committee put out a request for proposals to carry out an extensive public acceptance and feasibility survey. The study tested two technologies, retinal scanning and fingerprinting, the two most widely used in access control and law enforcement, respectively. Because the study's outcome will affect all 50 states, the FHA did not want the tests performed in any one state.

"So, Orkand was hired to do this cookoff, " explains Murata. " It's a two-part process involving data collection and data processing. Because the data collection would affect us in the field and the way we do business, it [had to] be accurate. The data collection is complete and now Orkand is doing the processing. "

This February, the project report went to the steering committee, which will pass it on to the FHA. The information will be used to establish a rule related to the use of biometrics in commercial driver's licenses. The project is on a short time line because the Truck and Bus Safety Regulatory Reform Act of 1988 requires the highway administration to complete the rule making by December 31, 1990. Funding for this accelerated demonstration program was included in the bill.

THE ORKAND CORPORATION IS AN information systems analysis and management consulting firm founded in 1970 and headquartered in Silver Spring, MD. Elizabeth Wenchel is Orkand's program director and manager of the California Personal Identifier Project. She has managed all of Orkand's past biometric projects, which include those done for the US Department of the Army and the FHA.

"Our research with biometric systems hasn't been related to secure access or necessarily related to security. We first got into it from a financial perspective, doing work with financial institutions and tracking some of the new technologies that were coming out," explains Wenchel.

Orkand's first project was with the Department of the Army, testing the feasibility of using an ATM with a hand geometry reader to distribute soldiers' pay. Hand geometry devices measure hand dimensions such as webbing and length and width of fingers. The identifier in this case was stored on a card instead of in a data base. A soldier went up to a machine, inserted his or her card, was verified, and received his or her pay from an ATM in cash.

"We tested the technology and determined it was feasible. However, the costs were quite exorbitant when compared to what your normal ATM transaction costs are. Adding a biometric identifier made the individual transaction cost too high," says Wenchel.

The firm's next project was for the Immigration and Naturalization Service (INS) to determine whether biometric technology could be used to cut down on green-card counterfeiting. Fraud is a major problem with the issue or reissue of green cards that have been lost or stolen. Orkand studied whether the immigration service should use an automated fingerprint identification system (AFIS). The major recommendation of that study was that INS use a more secure method of producing green cards, which the service adopted. Cards are now produced with a one-step machine, which is an improvement over the previously used multistep production method and eliminates any chance of intervention in card production. The study did not lead to the use of a biometric identifier but did bring about a more secure production method.

"In fact, the INS is now testing the use of biometrics' assistance at the border down in Southern California," Wenchel says." So they're continuing to look at that technology. We're not involved in that process. The highway administration was our next project. When the new legislation came out related to unique identification for commercial drivers, we met with the FHA and suggested that they may want to do some further study on it."

The DMV project is seven months long with California serving as the lead state for all other states. AAMVA, working from CDLIS, is developing a network called AAMVANET that will link all 50 states.

Wenchel explains, "If we establish that biometric identification is a feasible technology for the identification of commercial drivers, then those 50 states as represented by AAMVA would potentially add a biometric technology, link it onto their CDLIS system, and use the AAMVANET network for communication. They would all then be able to interface with each other to see whether a commercial driver has received a license in another state. So the project is directly related to commercial drivers, not to all drivers."

Should the highway administration develop rules for the use of biometric technology in CDLIS, the states will have an opportunity to respond to those rules and submit additional information. So, even if the Orkand report indicates that the technology is feasible, another time-consuming process must take place before any changes to the system occur.

Advances in image processing have allowed today's biometric improvements, according to Wenchel. Optical storage media allow high volumes of data to be stored cheaply in a small area. Chip technology advancements are also increasing the speed at which record searches can be conducted. These technological advances combine to make biometric systems better, cheaper, and faster.

Orkand is not testing whether the biometric technologies work, because they are already proven. Retinal scan technology is used in a number of secure areas throughout the country, such as missile silo sites, nuclear reactors, and information systems. Fingerprint technology is also used by a number of law enforcement agencies.

The biometric systems in place now are not the exact systems that could be used in a commercial driver's license environment, which requires a different function. Murata explains, "There are differences between access control technology people and law enforcement people. Access control security people are dealing with cooperative subjects. Those people want to be identified and want to have access. Law enforcement is dealing with people who don't want to be identifted. They have perfected what we call the cold search. " Law enforcement usually searches a data base for one match out of mainly possible choices. In access control, one-to-one comparisons are run to see if the current sample matches the one sample for that individual oil file.

Murata continues, These two technologies have developed quite differently and haven't had to come together. We need both for this project. We need one-to-one matching and rapid answers at times. We also need the law enforcement side to deter individuals from trying to create multiple identities. Unfortunately, there has been no marriage between these two different technologies. Both are very good at what they're focusing on, but there's been no pooling of the two. We'll know how successful they were in this environment following the tests. "

Orkand is testing the matching process to see if large numbers of people can be accurately identified. It is also testing, public acceptance. if people don't want to use the technology, the DMV probably will not implement it. The tests are also measuring the time required to capture a record in the driver's licensing environment. Because the DMV is customer oriented, a process that takes more than two minutes is too long.

The study itself was exhaustive. Researchers had to be sure the system was accurate and difficult to circumvent. Murata and his associates estimated 5 million to 7 million drivers have commercial licenses in the 50 states. Internal research and development statisticians in California and Texas determined 30,000 to 35,000 records had to be captured to obtain valid results about the system's accuracy and public acceptance.

Thirty-one thousand samples were collected for each technology. One thousand samples were collected as a control group. Volunteers were used to avoid invasion of privacy and security issues. Five thousand samples for each technology were duplicates, used to simulate the license renewal process. Volunteers giving duplicate samples used two capture devices-either two retinal scanners or two fingerprint readers-to test the consistency of samples taken from different machines. Using two machines also simulates situations where drivers who come in for renewal four years after giving their original samples have to use different data collection equipment.

In all. 26.000 unique records along with 5,000 duplicate samples were collected for each technology. Two types of collection took place. Some sites in California collected high volume samples and the duplicate samples, while other sites conducted public acceptance and timing surveys-three locations each in California, illinois, Texas, and North California. Each DMV office housed just one kind of technology, and DMV operators collected the data. Orkand helped on-site by recruiting volunteers for the tests and bringing them to the operators.

Explains Wenchel, They did the data collection, and that's where we did all the public acceptance surveys-in those 12 offices. We did all the timing Surveys in those 12 offices. We are doing studies on how long it took to train the DMV operators and how it fits in with their overall procedures. So in those 12 sites in those four states the testing was very. very controlled."

The DMVs made their personnel available for the duration of the project, which took from two to four weeks in each state. Orkand provided data collection personnel at additional sites in California. which collected the high volume data. The tests were still conducted in DMV offices and used volunteers.

Orkand collected data from August to December, and data preparation was done in December. January was spent analyzing all the timing surveys and public acceptance surveys and conducting DMV manager and operator surveys. System testing was also done in January in Beaverton, OR, and Tacoma, WA.

THE MACHINES USED FOR THE STUDY are the EyeDentification System 7.5 and Morpho's automated fingerprint identification system (AFIS). The retinal scan device uses a low-intensity infrared light source to scan the circular area of the retina and take 320 readings of the vascular pattern at the back of the eye. The system involves three processes: enrollment, PIN verification, and recognition.

During enrollment, the subject looks into the lens of a machine, focuses on a visual alignment target, and presses a button to begin the retinal scan. The eye scan takes less than a minute. During verification, the waveform of the live eye signature is matched with the stored eye signature labeled with the same personal identification number (PIN). Recognition compares the waveform of the live eye signature with all stored signatures.

A person is accepted or rejected depending on whether a match exists. When an eye signature is recognized, the system activates a release of whatever security mechanism is used, and an authorized person is granted access. Instructions for use are printed on each machine, and the total process time is less than seven seconds.

Fingerprints were collected two ways, Some were rolled onto cards using an inkless method and will be scanned into the computer system later. Other samples were collected with a direct reader, an electronic capture device-and added to the data base. Future fingerprinting checks on commercial drivers can be done using either method. DMVS with direct readers will take a sample, transmit it to a central data base for comparison, and receive an answer within a few minutes. Offices without a direct reader will take samples on cards and mall them in for comparison. Samples sent in for the first time will be compared to all samples in the data base-a one-to-many search. A sample that has been previously enrolled may be from a driver applying for a fraudulent license. Samples from drivers who come in for license renewals will be compared to their records already on file-a one-to-one search to verify identity.

Videotapes taken in the Sacramento, CA, DMV office show volunteers giving samples using both retinal scan and fingerprint technologies. Some volunteers had trouble focusing on the visual target in the retinal scan device, but the scanning still took only one minute. Both fingerprinting methods were demonstrated using volunteers. The inkless method of rolling fingerprints onto cards, with two samples taken of each index finger, took less than a minute. The direct reader method, taking samples of two fingers, took over two minutes. None of the methods is especially difficult or time-consuming. Vendors supplied data collection equipment for the study at no charge. North American Morpho Systems of Tacoma, WA, provided over 16 fingerprinting units, and Eyedentify Inc. of Beaverton, OR, provided eight retinal scan units. The companies shipped the units to the sites and between locations, when necessary, and provided all data storage tapes and diskettes. Rolled fingerprints were collected with an inkless method developed by Identicator Corporation of Marina del Ray, CA, which provided the cards for data collection.

Wenchel explains, "There's only one retinal scanning company to my knowledge, but there are a number of AFIS vendors in the US. All vendors who have biometric systems-either fingerprinting or retinal scan systems-were invited to participate in this project. Originally it was envisioned that there would be probably three fingerprint vendors participating in the project. As it turned out, Morpho was the one who finally agreed to participate, and the others, for one reason or another, decided they could not make the equipment available at that time. So Morpho is serving as a representative of fingerprinting technologies."

Although technology such as biometrics offers significant benefits to society, not everyone is welcoming the Star Trek era with open arms. Some view it more as the age of Big Brother. As Orkand's Dr. Schwartz points out, "Compiling data bases, taking people's fingerprint information or their retinal scan information, relating it, and putting it into some nationwide data base runs into real or perceived issues of privacy, confidentiality, Big Brother. It conjures up the worst."

Wenchel adds that CDLIS is "not really a national system. It's a system comprised of states that will have their own systems for the identification of commercial drivers. It's just that the repository will be available, to all other 50 states. It will be a central repository for the 50 states and information on their state drivers. From a systems standpoint it would be difficult to have a network where every state is going out to every other state to make inquiries."

According to Schwartz, problems with biometrics are not related to the technologies themselves. Americans don't mind putting their faces or fingers up to a machine, but they do want to know why an agency is collecting the information and what they're going to use it for. As Wenchel points out, "If people want to get multiple licenses and don't want to be identified, they're not going to want either one of these machines. If they're used to having multiple licenses and using false social security numbers and false identifications, they're going to have a problem with this identification system.

"Of course, your AFIS vendor will say, Well, nobody wants to use an eye device' and your retinal scan vendor will say, Nobody wants to use a fingerprint device.' So you really have to go the source, which is the public. It'll be really interesting to see what the public reaction is. We're collecting data not just from the people who volunteer to use the devices and participate in this project but also from the people who refuse to participate. So we think that the study is going to be invaluable."

Even as the age of science fiction becomes real life, what does the future have in store for biometric technology Will houses unlock and cars start wit the touch of a finger in the next few years?

"I think you're going to see more and more use of biometrics as it becomes more proven," says Wenchel. "I think this particular report will encourage a lot of people to take another look at biometrics and its applications. If one of these technologies is proven to be feasible, and certainly if it's used for driver's licensing for commercial drivers, you're going to see it in virtually every licensing office in the United States. But that's a lot of "ifs. "

Schwartz sees biometrics going far beyond the driver's licensing environment. I see two long-term trends in our society intersecting. The first is a seriously increased awareness of security - it's on everybody's mind. The second is technological advance in itself, and that's inexorable. When those two come together, you have technology driving the development of products for which societal changes are creating a market.

"I think security managers are going to be hearing more and more about this and going to be using it," continues Schwartz. "Problems will arise, and they're going to be overcome, as with anything else. I can't imagine a single security practice that doesn't have a set of complaints or caveats attached to it. Professional people, I'm sure, work those problems through and bring them down to some acceptable range so that you get the maximum security with the minimum inconvenience. I think it's inevitable that that's going to happen here. We have good technologies. They're proven. "

Schwartz adds, "A decade ago, biometrics was featured in a James Bond movie. I thought it was bizarre, futuristic. It seemed so far off. And now you've got colleges using biometrics in cafeterias as a way of checking to make sure that a student doesn't get too many lunches. Biometric technologies are becoming mainstream right now.

"Driver's licensing is a very familiar environment. This use is commercial, but that's going to expand into other industrial security uses, I'm confident. And security people are the ones leading this charge."
COPYRIGHT 1990 American Society for Industrial Security
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 1990 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Title Annotation:California Personal Identifier Project will show whether biometrics can stop commercial drivers from obtaining duplicate licenses
Author:Haines, Kimberly A.
Publication:Security Management
Date:Apr 1, 1990
Previous Article:Back to the future with biometrics.
Next Article:The pitch for security.

Related Articles
Government study demonstrates leading performance for facial technology.
Foiling I.D. thieves: identity theft is on the rise, and states are looking for ways to secure their citizens' vital records.
FCC Commissioner demands media reform.

Terms of use | Copyright © 2016 Farlex, Inc. | Feedback | For webmasters