AmbAu: Accurate and Robust Physical Layer Authentication for Ambient Backscatter Devices

Ambient backscatter communication (AmBC) has become increasingly popular for facilitating ubiquitous Internet of Things (IoT) due to its ultra-low-power consumption and energy-harvesting capabilities. However, the open nature of wireless channels and the resource-limited devices in AmBC systems expose them to threats such as identity impersonation and wireless spoofing attacks. Recently, physical layer authentication (PLA) offered a lightweight solution to address the above security challenges by utilizing inherent wireless properties, like channel fading and signal patterns, as unique fingerprints to identify legitimate devices and distinguish them from attackers. Despite its promise, current research still lacks an effective PLA mechanism that can accurately authenticate backscatter devices (BDs) and defend against spoofing attacks in AmBC systems. This paper presents AmbAu, a two-stage PLA scheme designed to authenticate BDs by leveraging the signal correlation between an ambient source and the BD. In the initialization stage, the verifier extracts a complex covariance matrix from the downlink signals between the ambient source and the BD. This matrix is later used during the authentication stage to verify the BD's identity and detect spoofing attempts. Through security analysis, we demonstrate AmbAu's resilience against replay, relay, and counterfeiting attacks. Simulations under various conditions demonstrate AmbAu's accuracy, efficiency, and superior performance when compared to traditional PLA schemes.