Anomalous behavior detection-based approach for authenticating smart home system users

Area of Science:

• Cybersecurity
• Internet of Things (IoT) Security
• Behavioral Anomaly Detection

Background:

• Smart home systems (SHSs) face security challenges with unauthorized access to Internet of Things (IoT) devices.
• Existing authentication methods may not adequately protect against sophisticated threats in dynamic SHS environments.

Purpose of the Study:

• To introduce Duenna, an innovative authentication framework designed to secure smart home systems (SHSs).
• To ensure that only legitimate users can operate Internet of Things (IoT) devices within an SHS environment.

Main Methods:

• Duenna employs a behavioral anomaly detection (BAD) approach, hypothesizing that legitimate users exhibit typical behavioral patterns.
• It operates in two modes: initialization to build anomaly detection (AD) models from historical user and system data, and interactive operation to detect and reject anomalous commands.

Main Results:

• Empirical evaluation on real-world data demonstrates Duenna's high authentication rates.
• The framework effectively distinguishes between legitimate and malicious user activities, ensuring both security and a positive user experience.

Conclusions:

• User behavior-based authentication is a promising security strategy for SHSs.
• Duenna offers a robust and adaptable security scheme that can be integrated into existing smart home platforms.