Database-based Physical Layer Authentication for Dynamic Underwater Acoustic Networks

Physical layer authentication (PLA) mechanisms play an important role on enhancing the security of underwater acoustic networks (UANs) in a lightweight and effective manner. However, the existing PLA mechanisms have been significantly affected by harsh underwater acoustic channels with severe temporal and spatial variations, which may inevitably lead to the PLA performance degradation in UANs. Therefore, we propose a database-based PLA scheme in this paper, with the aim of improving the authentication accuracy and robustness for dynamic UANs. A dynamic database that relies on the channel impulse response (CIR) is designed by fully tracking and exploiting the time-varying underwater acoustic links from neighbors. The time-varying resistant feature vectors are constructed through matching processes among CIR databases, which lay a significant foundation for reliable PLA in UANs. Through the construction of the dynamic CIR database using a updation mechanism, the proposed PLA scheme can learn and accommodate to the dynamic underwater environment and thus achieve a strong PLA performance in UANs. Simulation results and sea trial studies have shown the high authentication accuracy of the proposed PLA scheme in UANs.