Imperial College London researchers have refined an authentication system based on gait biometrics that they say meets National Institute of Standards and Technology and Dieharder standards.
As Tech Xplore reports, the system is based on multiple sensors positioned throughout the user’s body. Each of these sensors can capture movement patterns from its own particular position, but they’re all brought to together in a virtual framework of the user’s body. In this way, their signals can be combined to assess overall movement patterns in the subject.
The system was outlined in a recent paper published in IEEE Explore. Speaking to Tech Xplore’s Ingrid Fadelli, lead author Yingnan Sun said that the system is capable of “generating a 128-bit key within only 12 seconds of walking.” He added that the system is currently only based on acceleration signals, and that his team will seek to improve its performance by incorporating other kinds of gait signals such as gyroscopic data.
Given the system’s need for multiple sensors worn over a user’s body, its commercial applications are unclear; other gait recognition systems have been developed that are designed to identify a user with no need for a network of wearable sensors. Still, the use of a virtual model onto which sensor signals are projects – which the researchers call an ‘artificial neural network framework’ – could yield interesting insights and lead to some highly innovative biometric technologies as the research continues.
Source: Tech Xplore
August 22, 2018 – by Alex Perala