Walking could form part of biometric ID for wearables
By analyzing the gait of a person's walk, researchers from the CSIRO have created both a unique way to provide power to wearable devices and a new authentication method that one day may replace conventional security methods such as passwords or fingerprints.
Working at CSIRO's Data61 research unit, investigators wanted to overcome the current problems of high battery drain when using accelerometers to measure motion and velocity to electronically determine a person's gait. To do this, the team at Data61 turned to the technique of kinetic energy harvesting (similar to that used in some prototype footwear), to convert body movements into electrical energy and prolong battery charge.
However, rather than simply relying on accelerometers to help determine a person's individual walking style, the researchers found that energy generation derived from human movement actually produced distinctive patterns that could be used as another unique method of personal authentication.
"By applying both techniques we have developed a way to achieve two goals at once – powering devices and the ability to verify a person's identity using a wearable device by capturing the energy generated from the way they walk," said Data61 researcher Sara Khalifa.
Known as KEH-Gait, the technology was tested with users walking across a range of different surfaces and locations – including interior materials such as carpet, and exterior coverings, including grass and asphalt – to determine any changes in gait over periods of time and changes in walking surfaces.
According to the researchers, the results demonstrated that KEH-Gait is able to attain a 95-percent rate of authentication accuracy, while reducing energy drain by around 78 percent when compared to regular accelerometer-measured techniques of authentication.
The team at Data61 also claim that their KEH-Gait system could also generally spot "imposters" who tried to emulate someone else's gait, with just 13 out of 100 attempts being incorrectly accepted by the system as genuine. This leads the researchers to believe that their technique may one day be a useful adjunct – or even a replacement – for conventional methods of identification such as signatures, passwords, or fingerprints.
"Firstly, it is convenient because as we walk around each day our gait can be sampled continuously and verified without us having to manually adjust anything," said Professor Dali Kafaar, leader of the Networks Research Group at Data61. "Secondly, it's more secure than passwords because the way we walk is difficult to mimic. Since the KEH-gait keeps authenticating the user continuously, it collects a significant amount of information about our movements, making it difficult to imitate or hack unlike guessing passwords or pin codes."
With everything from electronically-assisted leg attachments to help you slice across the snow to devices that alert that an asthma attack is imminent, wearable devices are fast becoming a large global market that is forecast to pull in around US$150-billion by the year 2026. With this burgeoning trade, the Data61 researchers believe that their techniques could tap into a wide range of devices to both improve battery life and increase personal electronic security.
"With many of us already tracking our health using wearable devices there is a great opportunity to explore new authentication methods based on our movements," said Professor Kafaar.
Future research for the team will include investigating other higher security and continuous authentication methods in areas such as identifying and employing the specific differences in people's breathing patterns.
The results of this research were recently presented to the Internet Society (PDF).