On the Secrecy Design of STAR-RIS assisted Uplink NOMA Networks

Zheng Zhang, Jian Chen, Yuanwei Liu, Qingqing Wu, Bingtao He, Long Yang

Introduction

As the wireless industry moves toward 6G, technologies like Non-Orthogonal Multiple Access (NOMA) and Reconfigurable Intelligent Surfaces (RIS) are becoming central to achieving high spectral efficiency and programmable environments. However, conventional RIS is limited by its reflection-only nature, serving only users on one side of the surface. Simultaneously Transmitting and Reflecting RIS (STAR-RIS) overcomes this by allowing signals to both pass through and reflect, providing full 360° coverage.

While STAR-RIS enhances connectivity, it also expands the potential for security breaches. In an uplink scenario, signals intended for the base station can leak through or reflect toward an eavesdropper, regardless of their position relative to the surface. This paper proposes a novel security framework to maximize secrecy in STAR-RIS assisted uplink NOMA networks.

System Model and Problem Definition

The study considers an uplink network where an indoor user (IU) and an outdoor user (OU) transmit data to a base station (BS) via a STAR-RIS. An eavesdropper (E) attempts to intercept these transmissions.

System model of a STAR-RIS assisted uplink NOMA network with an eavesdropper.

The primary challenge is the "opposite-side leakage" unique to STAR-RIS. Unlike traditional RIS, where an eavesdropper must be on the same side as the users to intercept reflected signals, STAR-RIS allows signals to leak to the opposite side via transmission. The authors address two scenarios:

  1. Full Eavesdropping CSI: The exact channel state information (CSI) of the eavesdropper is known, allowing for adaptive-rate wiretap coding to maximize the minimum secrecy capacity.
  2. Statistical Eavesdropping CSI: Only the average channel behavior is known, necessitating constant-rate wiretap coding to minimize the maximum Secrecy Outage Probability (SOP).

Methodology: Alternating Hybrid Beamforming

To tackle the non-convex optimization of beamforming and power allocation, the authors developed an Alternating Hybrid Beamforming (AHB) algorithm. The algorithm decouples the complex variables into manageable sub-problems:

  • Passive Beamforming: Utilizing a "polarization identity" technique, the STAR-RIS coefficients are optimized iteratively to align signals at the BS while nulling them at the eavesdropper.
  • Power Allocation: Optimal power levels are derived in closed-form, ensuring computational efficiency.
  • Rank-One Constraints: Difference-of-Convex (DC) relaxation is applied to ensure the beamforming matrices are physically realizable.

The performance was validated using Monte-Carlo simulations with Rician and Rayleigh fading models, averaging results over 100 random channel realizations.

Key Insights and Results

The research yields several critical findings for the design of secure 6G networks:

  • Quantization Requirements: While 3-bit quantization of phase shifts is often sufficient for throughput, secrecy performance requires at least 4-bit quantization. This finer control is necessary to accurately suppress signals at the eavesdropper's location.
  • Deployment Strategies: The optimal location for the STAR-RIS depends on the available CSI. With full CSI, placing the surface near the users or BS is ideal. With only statistical CSI, a more distant deployment helps mitigate the eavesdropper's potential gains.
  • NOMA Advantage: NOMA consistently outperforms Orthogonal Multiple Access (OMA) by leveraging the extra degrees of freedom provided by STAR-RIS for better interference management.
  • Efficiency: The AHB algorithm demonstrates rapid convergence (50–100 iterations), suggesting feasibility for real-time implementation.

Conclusion and Future Work

STAR-RIS significantly enhances the secrecy performance of uplink NOMA networks compared to conventional RIS. However, the reliance on perfect legitimate CSI remains a limitation. Future research should focus on robust beamforming designs for high-mobility scenarios and extending the framework to handle multiple or colluding eavesdroppers.

Summary updated

BibTeX
@article{zhang2022secrecy,
  author = {Zheng Zhang and Jian Chen and Yuanwei Liu and Qingqing Wu and Bingtao He and Long Yang},
  title = {{On the Secrecy Design of STAR-RIS assisted Uplink NOMA Networks}},
  journal = {IEEE Trans. Wireless Communications},
  year = {2022},
  doi = {10.1109/twc.2022.3190563},
}
Figure