CONTINUOUS AUTHENTICATION FOR MODERN PERSONAL DEVICES

Abstract

There is a rapid increase in the use of personal computing devices and various security-critical services are available on these devices. Therefore, Continuous Authentication (CA) has gained popularity as a suitable authentication scheme for personal devices. We identify and address concerns preventing widespread adoption of CA. We focus on; (1) additional resource consumption and (2) privacy invasion, of CA. The generated Resource Profile Curve (RPC) for biometrics reveal the trade-off between utility and resource usage. RPCs highlight IMU-gait as a suitable biometric for CA. The privacy invasiveness of IMU-gait was studied to show that gait data can reveal physical attributes with high accuracy. We highlight the drastic variance in gait when the actions change. We propose action invariant gait features to provide robust performance with a single action enrolment. This thesis shows that a CA system utilizing gait can overcome common concerns and facilitate widespread adoption with single action enrolment.