Simultaneous Transmitting and Reflecting (STAR)-RIS for Harmonious Millimeter Wave Spectrum Sharing

Omar Hashash, Walid Saad, Mohammadreza F. Imani, David R. Smith

Introduction

The millimeter-wave (mmWave) spectrum is a critical frontier for 5G and Beyond 5G (B5G) networks, offering the massive bandwidth necessary for high-data-rate applications. However, the deployment of mmWave technology faces two primary hurdles: severe signal blockage due to high path loss and the presence of incumbent services, such as radar systems, that already occupy these frequency bands. To facilitate the coexistence of secondary 5G networks and primary incumbent services, efficient spectrum sharing is essential.

Traditionally, Reconfigurable Intelligent Surfaces (RIS) have been employed to mitigate blockages by creating virtual line-of-sight (LoS) paths. However, conventional RIS are limited to reflecting signals, which often forces a performance compromise when trying to simultaneously enhance a user's signal and nullify interference toward a primary receiver. This paper proposes a novel solution using Simultaneous Transmitting and Reflecting Reconfigurable Intelligent Surfaces (STAR-RIS) to achieve harmonious spectrum sharing.

The STAR-RIS Paradigm

Unlike conventional RIS, which only operate in a reflection-only mode, STAR-RIS elements can simultaneously transmit and reflect incident signals. This capability allows the surface to provide a full 360-degree coverage area, effectively dividing the environment into a transmission region and a reflection region.

In the context of spectrum sharing, this spatial decoupling is transformative. By placing the primary network (e.g., a radar receiver) in the transmission region and the secondary 5G users in the reflection region, the system gains additional degrees of freedom. The STAR-RIS can independently control the phase shifts for both regions, allowing it to boost the secondary network's performance while precisely managing interference toward the primary network.

Methodology and Optimization

The researchers modeled a system featuring a multi-antenna Base Station (BS), multiple secondary users, and a primary network pair. The STAR-RIS operates in Energy Splitting (ES) mode, where the energy of each incident signal is divided into transmitted and reflected components. While this results in a lower amplitude for each component compared to a full-reflection RIS, the independent phase control compensates for the power reduction.

To maximize the system's efficiency, the authors formulated a joint optimization problem aimed at maximizing the downlink sum-rate of the secondary network. This optimization is subject to a critical constraint: the primary network must maintain a minimum Signal-to-Interference-plus-Noise Ratio (SINR).

Because the resulting mathematical problem is non-convex and highly complex, the authors employed an Alternating Iterative Algorithm. This approach involves:

  1. Beamforming Optimization: Fixing the RIS coefficients to solve for the BS beamforming matrix.
  2. Coefficient Optimization: Fixing the beamforming matrix to solve for the STAR-RIS transmission and reflection coefficients using Successive Convex Approximation (SCA).

This iterative process continues until the algorithm converges on a high-performing, sub-optimal solution.

Performance Evaluation

The proposed framework was evaluated through numerical simulations based on 3GPP-compliant mmWave models at 28 GHz. The setup included a BS with 16 antennas, 4 secondary users, and a STAR-RIS with up to 32 elements.

Key Results:

  • Spectral Efficiency Gains: The STAR-RIS scheme achieved a 14.57% improvement in spectral efficiency compared to conventional "adjacent" RIS setups (where separate surfaces are used for reflection and transmission).
  • Peak Throughput: The system reached a spectral efficiency of 25.87 bits/s/Hz with 32 RIS elements.
  • Scalability: Performance improved significantly with the addition of more RIS elements ($N$) and increased Base Station power ($P_{max}$). Notably, doubling the BS power resulted in a ~70% gain, while doubling the number of RIS elements yielded a ~52% gain.
  • Interference Management: The system demonstrated high sensitivity to the primary network's requirements. As the interference threshold for the radar became stricter, the algorithm dynamically biased more energy toward interference mitigation, ensuring the "harmonious" coexistence promised in the title.

Discussion and Limitations

While STAR-RIS provides a significant leap in flexibility, it introduces hardware and operational complexities. The study highlights the trade-off between the "power hit" taken during energy splitting and the gains provided by independent phase control.

Several challenges remain for real-world implementation:

  • Channel State Information (CSI): The current model assumes perfect CSI. In practical mmWave environments, accurately estimating the channels for both transmission and reflection regions is a significant hurdle.
  • Mobility: The simulation assumes a static environment. Future research must address how STAR-RIS handles high-mobility users, such as vehicles or drones.
  • Hardware Cost: The elements required for independent phase control in both transmission and reflection are more complex to manufacture than standard reflect-only elements.

Conclusion

This research demonstrates that STAR-RIS is a powerful tool for solving the spectrum scarcity problem in mmWave bands. By decoupling the tasks of signal enhancement and interference nulling, STAR-RIS enables 5G networks to operate efficiently alongside incumbent services, paving the way for more dense and versatile wireless ecosystems.

Summary updated

Abstract

The opening of the millimeter wave (mmWave) spectrum bands for 5G communications has motivated the need for novel spectrum sharing solutions at these high frequencies. In fact, reconfigurable intelligent surfaces (RISs) have recently emerged to enable spectrum sharing while enhancing the incumbents' quality-of-service (QoS). Nonetheless, co-existence over mmWave bands remains persistently challenging due to their unfavorable propagation characteristics. Hence, initiating mmWave spectrum sharing requires the RIS to further assist in improving the QoS over mmWave bands without jeopardizing spectrum sharing demands. In this paper, a novel simultaneous transmitting and reflecting RIS (STAR-RIS)-aided solution to enable mmWave spectrum sharing is proposed. In particular, the transmitting and reflecting abilities of the STAR-RIS are leveraged to tackle the mmWave spectrum sharing and QoS requirements separately. The STAR-RIS-enabled spectrum sharing problem between a primary network (e.g. a radar transmit-receive pair) and a secondary network is formulated as an optimization problem whose goal is to maximize the downlink sum-rate over a secondary multiple-input-single-output (MISO) network, while limiting interference over a primary network. Moreover, the STAR-RIS response coefficients and beamforming matrix in the secondary network are jointly optimized. To solve this non-convex problem, an alternating iterative algorithm is employed, where the STAR-RIS response coefficients and beamforming matrix are obtained using the successive convex approximation method. Simulation results show that the proposed solution outperforms conventional RIS schemes for mmWave spectrum sharing by achieving a 14.57% spectral efficiency gain.
BibTeX
@inproceedings{hashash2023simultaneous,
  author = {Omar Hashash and Walid Saad and Mohammadreza F. Imani and David R. Smith},
  title = {{Simultaneous Transmitting and Reflecting (STAR)-RIS for Harmonious Millimeter Wave Spectrum Sharing}},
  booktitle = {WCNC},
  year = {2023},
  doi = {10.1109/wcnc55385.2023.10118966},
}
Figure