Secrecy Analysis of CSI Ratio-Based Transmitter Selection with Unreliable Backhaul

TL;DR

This work tackles secrecy in a multi-transmitter system with unreliable wireless backhaul by proposing a CSI ratio-based transmitter selection (RTS) that maximizes the destination-to-eavesdropper channel power-gain ratio. Backhaul unreliability is modeled with a mixture distribution incorporating a reliability factor Δ, and NZR and SOP are derived for scenarios where backhaul activity is either known or unknown. The authors derive closed-form expressions and asymptotic results, showing that RTS can approach the performance of the optimal scheme without requiring noise power or secrecy-rate evaluation, with improvements when backhaul knowledge is available. Practically, the RTS strategy offers simple, robust secrecy enhancement in heterogeneous networks, particularly when backhaul activity information is available, and the secrecy gains scale with the number of transmitters K and the reliability Δ.

Abstract

This paper explores the secrecy performance of a multi-transmitter system with unreliable backhaul links. To improve secrecy, we propose a novel transmitter selection (TS) scheme that selects a transmitter with the maximum ratio of the destination channel power gain to the eavesdropping channel power gain. The backhaul reliability factor is incorporated with the distribution of the channel power gain through the utilization of a mixture distribution. We evaluate the non-zero secrecy rate (NZR) and the secrecy outage probability (SOP) as well as their asymptotes in two scenarios of backhaul activity knowledge, where it is available and where it is unavailable. The results illustrate that because of the unreliable backhaul, the proposed destination-to-eavesdropper channel power gain ratio-based TS scheme is constrained in terms of secrecy performance. However, performance enhancements are observed when the backhaul knowledge activity is utilized. Furthermore, the proposed scheme outperforms all the sub-optimal TS schemes and achieves nearly optimal performance without requiring noise power or the evaluation of the exact secrecy rate measurement.

Figures (4)

• Figure 1: A small-cell network with backhaul links.
• Figure 2: The NZR versus $1/\lambda_{D}$ varying $K$ and $\Delta$.
• Figure 3: SOP versus $1/\lambda_{D}$ varying $K$ and $\Delta$.
• Figure 4: SOP comparison of different transmitter selection schemes.