Scientists Make Energy-Generating Keyboard That Knows Who’s Typing On It

Passwords continue to be a glaring weakness in digital security. And while biometric alternatives, such as fingerprint readers, are finding their way onto more consumer electronics devices, they are not without their limitations either. So what about tightening the security screw further by applying a continuous biometric — such as a keyboard that knows who’s typing on it in real-time and could lock out an intruder based on sensing unfamiliar keystrokes?

A team of scientists including from the Georgia Institute of Tech and research labs in China have developed just such a prototype device. As well as sensing touch and typing style, the non-mechanical keyboard is able to generate energy to power itself, and self cleans, thanks to a dirt repellent coating and crafty composition — thin layers of plastic, topped and tailed with conductive film — so there’s no mechanical gullies to harvest the crumbs of your lunch. The low profile keys have a top electrification layer that allows them to generate charge from typing friction (aka contact electrification) — so the keyboard is able to register individual keystroke data and track typing style, or to harness typing friction to generate electrical charge for powering other devices.

“The keyboard identify the force and speed at which the typer input characters. This is not available in today’s cell phone at all. The keyboard can identify personality in information input so that it is a high level of security,” Zhong Lin Wang, one of the scientists involved in the research, told TechCrunch. “The power generated can automatically send signals to the computer so that no power is needed for the keyboard.”

In a paper detailing the research project, called Personalized Keystroke Dynamics for Self-Powered Human–Machine Interfacing — published in the ACSNano journal — the scientists argue that algorithms which can identify who is typing by analyzing keystroke timing are more limited than their hardware-based approach of “typing-induced electric signals”, as the latter is also able to sense pressure, as well as speed and strike timing.

They write:

…keystroke dynamics has developed rather slowly and is still at its very early stage, partially because almost all of the proposed studies rely on dimensional keystroke timing vectors as typing patterns, which can only communicate the keystroke timing characteristics, thus rendering this biometric measure lacking in universality, uniqueness, permanence, accuracy, and acceptability… A breakthrough is desperately needed to obtain a unique and permanent typing pattern that can be used as an effective identifier/calibration for practical applications. In this work, the typing-induced electric signals can not only differentiate keystroke timing but also quantitatively record concrete dynamic changes in the course of typing using the self-generated electric current and voltage signals. It offers an unprecedentedly accurate, unique, and permanent typing pattern for further verification and recognition purposes.
First, through integration of the IKB [Intelligent Keyboard] with a signal-processing circuit, a complete touch-sensing system was built up. A wireless alarm is triggered once a finger gently taps on a key. Furthermore, each key is individually addressable, which enables tracing and recording the typing content in real time. This capability will provide administrators with great convenience for identifying impostors or intruders, and it will have extensive applications in keyboard-based information security.

No external power source is required for the keyboard to function, thanks to its self-powering abilities. So it is evidently a more energy-efficient alternative to tracking typing via capacitive data that could be harvested from typing on a touchscreen keyboard.

How much energy can be generated by typing on the smart keyboard? Enough to charge “small electronics at arbitrary typing speed greater than 100 characters per min”, according to the team — which they characterize as “a giant leap compared to previous reports”.

They add:

Given such features as self-securing, self-powering, and self-cleaning, as well as unique applicability resulting from distinctive mechanism and one-piece non-mechanical-punching structure, the IKB presented in this work is a practical approach in converting typing motions for self-powered electronics as well as a milestone in the development toward a highly secure behavioral biometrics-based authentication system, which will have extensive applications in the fields of artificial intelligence, cyber security, and computer or network access control.

The U.S. Department of Energy contributed funding to the research. For more details of the research read the full paper in ACSNano.

“The technology is ready for commercialization within a couple of years if some one is willing to invest,” added Zhong.