Towards a Unified Trust Framework for Detecting IoT Device Attacks in Smart Homes

Abstract

Trust in Smart Home (SH) Internet of Things (IoT) technologies is a primary concern for consumers, which is preventing the widespread adoption of smart home services. Additionally, the variety of IoT devices and cyber attacks make it hard to build a generic attack detection framework for smart home IoT devices. In this paper, we present a roadmap towards building a unified approach towards establishing trust scores as an indicator of the security status of an IoT device in a smart home that works across multiple attacks and device types/protocols. Specifically, we first introduce artificial reasoning inspired evidence collection approach by introducing a small set of factors that are affected significantly if a smart home IoT device is under attack. Thereafter, we propose an explainable trust scoring model that maps the device level evidence into trust scores in a way that produces lower trust scores when devices are under attack. Specifically, the trust model involves an Augmented Bayesian Belief based Model embedded with novel non-linear weighing functions; explicitly designed to account for the severity of the attack, probabilistic discounting of parts of the evidence caused by benign changes, thus explaining our success. For evaluation of the framework, we use two real datasets that contain a variety of actual cyber-attacks and benign traffic from seven different smart home IoT devices. Our evaluation seeks to investigate the generality of our framework across multiple datasets, with various classes of IoT devices and cyber attacks.