Root Cause Analysis of Anomaly in Smart Homes Through Device Interaction Graph

In smart homes, as a hallmark application of the Internet of Things (IoT), home automation platforms manage IoT devices via automation rules deployed by users, facilitating complex interactions for scenario-based services. However, maintaining and debugging such systems becomes challenging due to the potential for attackers to leverage complex device correlations for indirect attacks. We introduce a new practical approach for detecting the root cause of anomalies in home automation platforms. The key idea behind our approach is that causal dependencies dominate device interactions despite their complexity. Our approach offers insights into the root causes unattainable with traditional methods, hindered by gaps in knowledge and the limitations of manual inspections. Our evaluation of a synthetic dataset demonstrates that the proposed approach, aletheia, achieves comprehensive and precise root cause analysis for indirect attacks, with an average accuracy of 99.07%, significantly outperforms the state-of-the-art methods.