CSI‐based authentication: Extracting stable features using deep neural networks

Abstract

The first step of secure communication is authenticating users and detecting malicious ones. In recent years, some promising schemes have been proposed using wireless medium network's features, in particular, channel state information (CSI) as a means of authentication. These schemes mainly compare the user's previous CSI with the new received CSI to determine if the user is what it is claiming to be. Despite high accuracy, these approaches lack the stability in authentication when the users rotate in their positions. It is due to the significant change in CSI when a user rotates, which misleads the authenticator when it compares the new CSI with the previous ones. Our approach presents a way of extracting features from raw CSI measurements, which are stable with respect to rotation. We employ a deep neural network to extract these features. We also present a scenario in which users can be efficiently authenticated while they are at specific locations in an environment (even if they rotate) and will be rejected if they change their location. In addition, experimental results are presented to show the performance of the proposed scheme.