Performance Analysis and IRS Elements Allocation for Active IRS-assisted Secure Networks
Sihao Lai, Yan Wang, Jiatong Bai, Xuehui Wang, Feng Shu · 2025
In this paper, an active intelligent reflective surface (IRS) is divided into two sets for reflecting confidential messages (CM) and artificial noise (AN), respectively. Fixing the total number of active IRS elements, the element allocation ratio $\alpha$ of two sets is varied to analyze the secure rate (SR) of active IRSassisted secure network under the Rayleigh channel. Using Nullspace projection (NSP), equal-gain reflection (EGR), and the law of large numbers, a closed-form (CF) expression for the SR is derived. It is found that when the direct path channel exponents increase, as $\alpha$ increases, the SR first decreases and then increases. Then, the impact of the amplified noise power at the active IRS on the SR is analyzed. With increasing $\sigma_{i}^{2}$, regardless of $\alpha$, the SR is convergent.