Average Secrecy Capacity Maximization of Rotatable Antenna-Assisted Secure Communications

TL;DR

This paper proposes a method to maximize the average secrecy capacity in wireless systems using rotatable antennas, deriving a closed-form near optimal deflection angle and analyzing secrecy outage probabilities.

Contribution

It introduces a closed-form near optimal deflection angle solution for secrecy capacity maximization with rotatable antennas under practical constraints.

Findings

The objective function is quasi-concave with respect to the adjustment factor.

The closed-form solution closely matches the optimal solution in simulations.

Secrecy outage probabilities align with theoretical predictions at high SNR.

Abstract

A rotatable antenna, which is able to dynamically adjust its deflection angle, is promising to achieve better physical layer security performance for wireless communications. In this paper, considering practical scenarios with non-real-time rotatable antenna adjustment, we investigate the average secrecy rate maximization problem of a rotatable antenna-assisted secure communication system. We theoretically prove that the objective function of the average secrecy rate maximization problem is quasi-concave with respect to an adjustment factor of the rotatable antenna. Under this condition, the optimal solution can be found by the bisection search. Furthermore, we derive the closed-form optimal deflection angle for the secrecy capacity maximization problem, considering the existence of only line-of-sight components of wireless channels. This solution serves as a near optimal solution to the average secrecy rate maximization problem. Based on the closed-form near optimal solution, we obtain the system secrecy outage probability at high signal-to-noise ratio (SNR). It is shown through simulation results that the near optimal solution achieves almost the same average secrecy capacity as the optimal solution. It is also found that at high SNR, the theoretical secrecy outage probabilities match the simulation ones.