It seems like there’s a touchscreen on everything these days: your phone, the checkout kiosk at the grocery store, your car. Sure, they’re handy. You can input a PIN, a password, or even sign your name. But maybe some things just shouldn’t have them. Before you mar your nice wooden door with a touchpad controlled lock, consider what engineers at Rutgers University have done.
Led by Yingying Chen, a professor of electrical and computer engineering at Rutgers, they created a biometric access system, called VibWrite, that should work on any solid surface—no touchpad or fingerprint sensor required. The system reads the change that finger pressure and motions make in the vibrations on say a door or a tabletop or, Chen expects, any other solid object.
When tested on a wood door and a wood table VibWrite verified identities with greater than 95 percent accuracy and with a false-positive rate of less than 3 percent. Chen delivered the details of the technology this week at the ACM Conference on Computer and Communications Security, in Dallas, Tex.
Images: The DAISY Lab
The system was made up of a vibration motor, a vibration sensor, and a computer. (The computer could be replaced with a fairly low-end CPU capable of running some signal processing and machine learning, says Chen.) The motor sent low-amplitude vibrations at 20 kHz—too high a pitch for humans to feel—through the wood. The presence or absence of a finger touching the wood changed the vibrations felt at the sensor in a way that the CPU was able use to identify people. That vibrational change is both a physiological signature—the result of the shape of your finger and the mechanical properties of its bones—and a behavioral one—how hard you press when making particular motions.
“The eventual goal is to sign your name on the door” and have it unlock, says Chen. But at the moment, the system can only recognizes simple symbols, such as the Greek letters pi and alpha. It can also do the equivalent of a PIN number or the lock pattern on an Android phone.
Right now, the system can pick up motion an area 20-50 centimeters from the sensor, but that distance is a function of cost, says Chen. “Currently we’re using a very simple motor and receiver,” she says. “We were trying to contain costs below $50. But a better motor and receiver give better accuracy and larger distances.”
Chen says VibWrite could be commercialized in “a couple of years.” But first her team needs to improve its false negative rate and test it on more surfaces and in different conditions. That’ll be particularly important if you ever plan to unlock your car in the rain by writing “Open Sesame” on the window with your finger.
This post was corrected on 12 November 2017 to indicate that it can recognize more than two symbols.
Samuel K. Moore is the senior editor at IEEE Spectrum in charge of semiconductors coverage. An IEEE member, he has a bachelor's degree in biomedical engineering from Brown University and a master's degree in journalism from New York University.