On the Applicability of Users' Operation-action Characteristics for the Continuous Authentication in IIoT Scenarios
Guozhu Zhao, Pinchang Zhang, Xiaohong Jiang · 2020
Industrial Internet of Things (IIoT) systems are facing an urgent demand for continuously protecting sensitive information during the manufacturing processes. This paper studies the problem of user authentication across various scenarios in IIoT systems, based on operation features. Specifically, utilizing the routine operations (e.g., walking, scanning, and touchscreen), a user authentication approach is proposed to continuously and implicitly validate the identity of current user, and the availability and reliability of operation features are investigated. A systematic authentication method is then designed to evaluate the distinctiveness and accuracy of operation characteristics. With the help of support vector machine (SVM), a two-class classifier is also developed to discriminate between legitimate users and impostors. Finally, we construct extensive experiments to verify the discriminability and reliability of the proposed authentication approach.