Enhancing Performance and Security of IRS-NOMA Systems with Multiple Eavesdroppers and Hardware Impairments

초록

This article examines the integration of non-orthogonal multiple access (NOMA) with intelligent reconfigurable surfaces (IRS). It focuses on addressing these challenges while mitigating the adverse effects of residual hardware impairments (RHI), which cause signal distortion. The primary contribution of this paper lies in deriving precise closed-form expressions for both connection outage probability (COP) and intercept probability (IP) within the proposed system. Additionally, a Deep Neural Network (DNN) model was developed to estimate COP and IP, complemented by Monte Carlo simulations to validate analytical results. Simulation findings reveal clear quantitative performance gains. For instance, increasing IRS reflecting elements from 10 to 30 reduces the COP by 18% at a signal-to-noise ratio (SNR) of 20 dB. Similarly, increasing the amplitude reflection coefficient from 0.4 to 0.9 results in a 30% IP reduction. Increasing the number of eavesdroppers (Es) from 1 to 5 yields an IP reduction of 8% at an SNR of 15 dB. Similarly, lowering the hardware-impairment coefficient from 0.02 to 0.03 improves the COP by 17% at an SNR of 20 dB, while the IP decreases by 4% under the same conditions. Moreover, increasing the power allocation of the user from 0.2 to 0.3 yields a COP reduction of 4% at an SNR of 17 dB, while the IP decreases by 4% under the same conditions. Overall, the proposed IRSassisted NOMA architecture demonstrates substantial resilience and security improvements for Internet of Things (IoT) networks affected by hardware imperfections. © 2026 ADVANCES IN ELECTRICAL AND ELECTRONIC ENGINEERING.

키워드