Physical Layer Security of a Dual-Hop Regenerative Mixed RF/UOW System

Ensuring physical layer security is a crucial task in conventional and emerging communication systems, which are typically characterized by stringent quality of service and security requirements. This also accounts for wireless technologies in the context of the Internet of Things paradigm, which are expected to exhibit considerably increased computational complexity. Based on this, the present contribution investigates the secrecy outage performance of a dual-hop decode-and-forward (DF) mixed radio-frequency/underwater optical wireless communication (RF/UOWC) system. Such wireless network configurations are particularly useful in efficient and demanding scenarios, such as military communications. Therefore, our analysis considers one single-antenna source node (SS) communicating with one legitimate destination node (DD) via a DF relay node (RR) equipped with multiple antennas for reception. Particularly, the relay receives the incoming signal from SS via an RF link, applies selection-combining (SC) technique, fully decodes it, re-encodes it, and then forwards it to the destination via a UOWC link. The communication is performed under the eavesdropper's attempt to intercept the S-RS-R hop (RF side). In this context, a closed-form expression for the secrecy outage probability is derived along with a thorough asymptotic analysis in the high SNR regime, based on which the achievable diversity order is provided. The offered results provide useful insights on the impact of some key system and channel parameters on the secrecy outage performance, such as the number of eavesdroppers, the number of relay antennas, fading severity parameters of RF links, and water turbulence severity of the UOWC link. The conducted analysis shows that the secrecy outage probability is dominated only by the RR-DD link in the high SNR regime, regardless of the SS-RR parameters, such as the number of relay antennas and the average SNR at the relay branches. The offered analytic results are corroborated with respective results from computer simulations. Since these parameters are closely related with the computational complexity at the involved terminals, the offered insights are useful for the design and computationally sustainable operation of such systems.