Underwater acoustic communications are becoming a popular solution for underwater data communications and telemetry, making the authentication of transmitted data a necessity. In this paper, we propose a physical-layer authentication strategy for underwater acoustic networks (UWANs) with mobile devices. Such a scenario is more challenging than classical authentication scenarios in static networks, because the mobility of the receiver and/or transmitter implies that channel conditions slowly change over time. Thus, we cannot rely on the statistics of channel features to be stationary. In our proposed strategy, we assume that the receiver can rely on a set of sensors. We first extract a set of channel features, to be used to track the channel evolution over time. We then develop a long short-term memory (LSTM)-based approach, where at each step the sensors predict future feature values based on a learned model and on previously observed feature values. Next, each sensor computes the prediction error and passes it on to the actual receiver, which makes a decision on the signal authenticity through a generalized likelihood ratio test (GLRT). We model different classes of attacks and test them using simulation data obtained via the Bellhop ray tracing software. Numerical results show that our authentication mechanism successfully distinguishes between legitimate and impersonating transmitters, even when considering challenging attacking scenarios where the attacker can successfully mimic the channels between the legitimate transmitter and the sensors.

A RNN-based approach to physical layer authentication in underwater acoustic networks with mobile devices / Ardizzon, Francesco; Casari, Paolo; Tomasin, Stefano. - In: COMPUTER NETWORKS. - ISSN 1389-1286. - 243:(2024), pp. 11031101-11031112. [10.1016/j.comnet.2024.110311]

A RNN-based approach to physical layer authentication in underwater acoustic networks with mobile devices

Casari, Paolo;
2024-01-01

Abstract

Underwater acoustic communications are becoming a popular solution for underwater data communications and telemetry, making the authentication of transmitted data a necessity. In this paper, we propose a physical-layer authentication strategy for underwater acoustic networks (UWANs) with mobile devices. Such a scenario is more challenging than classical authentication scenarios in static networks, because the mobility of the receiver and/or transmitter implies that channel conditions slowly change over time. Thus, we cannot rely on the statistics of channel features to be stationary. In our proposed strategy, we assume that the receiver can rely on a set of sensors. We first extract a set of channel features, to be used to track the channel evolution over time. We then develop a long short-term memory (LSTM)-based approach, where at each step the sensors predict future feature values based on a learned model and on previously observed feature values. Next, each sensor computes the prediction error and passes it on to the actual receiver, which makes a decision on the signal authenticity through a generalized likelihood ratio test (GLRT). We model different classes of attacks and test them using simulation data obtained via the Bellhop ray tracing software. Numerical results show that our authentication mechanism successfully distinguishes between legitimate and impersonating transmitters, even when considering challenging attacking scenarios where the attacker can successfully mimic the channels between the legitimate transmitter and the sensors.
2024
Ardizzon, Francesco; Casari, Paolo; Tomasin, Stefano
A RNN-based approach to physical layer authentication in underwater acoustic networks with mobile devices / Ardizzon, Francesco; Casari, Paolo; Tomasin, Stefano. - In: COMPUTER NETWORKS. - ISSN 1389-1286. - 243:(2024), pp. 11031101-11031112. [10.1016/j.comnet.2024.110311]
File in questo prodotto:
File Dimensione Formato  
RNN-based-authentication-Underwater_2024_casari.pdf

accesso aperto

Tipologia: Versione editoriale (Publisher’s layout)
Licenza: Creative commons
Dimensione 1.48 MB
Formato Adobe PDF
1.48 MB Adobe PDF Visualizza/Apri

I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione

Utilizza questo identificativo per citare o creare un link a questo documento: https://hdl.handle.net/11572/405029
Citazioni
  • ???jsp.display-item.citation.pmc??? ND
  • Scopus 5
  • ???jsp.display-item.citation.isi??? 3
  • OpenAlex ND
social impact