No abstract

In the last years, smart-shoes moved from the medical domain, where they are used to collect gait-related data during rehabilitation or in case of pathologies, to the every-day life of an increasing number of people. In this paper, a method useful to effortlessly authenticate the user during gait periods is proposed. The method relies on the information collected by shoe-mounted accelerometers and gyroscopes, and on the distance between feet collected by Ultra-WideBand (UWB) transceivers. Experimental results show that a balanced accuracy equal to 97% can be achieved even when information about the possible impostors is not known in advance. The contribution of the different information sources, accelerometer, gyroscope, and UWB, is also evaluated.

No abstract