Deep Learning-based Continuous Authentication for an IoT-enabled healthcare service

Abstract

The Internet of Things (IoT) has introduced a new dimension to the Internet in the last decade; nonetheless, security, particularly attacks on authentication, continue to be a significant concern in IoT. The majority of research endeavours consider external attacks that originate from outside of an IoT network. Their authentication mechanisms authenticate users at the outset of a session. However, a device or user within the network may be a more significant threat than the external attacker due to their accessibility. An intruder during the session can physically grasp any IoT device and impersonate it. Therefore, the suggested security system continuously authenticates legitimate users inside a session. The system takes data from users and authenticates them using a Deep Learning-based Long Short-Term Memory classification algorithm. There are 3.5 percent false acceptances and 2.4% false rejections for the security system. The research also compared the suggested approach to other current security techniques.