Gait-Based Identification Using Wearables in the Personal Fog

Abstract

Wearables are becoming more computationally powerful, with increased sensing and control capabilities, creating a need for accurate user authentication. Greater control and power allow wearables to become part of a personal fog system, but introduces new attack vectors. An attacker that steals a wearable can gain access to stored personal data on the wearable. However, the new computational power can also be employed to safeguard use through more secure authentication. The wearables themselves can now perform authentication. In this paper, we use gait identification for increased authentication when potentially harmful commands are requested. We show how the relying on the processing and storage inherent in the personal fog allows distributed storage of information about the gait of the wearer and the ability to fully process this data for user authentication locally at the edge. While gait-based authentication has been examined before, we show an additional, low-power method of verification for wearables.