Spectral and Energy Efficiency of IRS-Assisted MISO Communication With Hardware Impairments

Shaoqing Zhou, Wei Xu, Kezhi Wang, Marco Di Renzo, Mohamed-Slim Alouini

Introduction

Intelligent Reflecting Surfaces (IRS) are a cornerstone technology for future 5G-Advanced and 6G networks. By utilizing a planar array of passive elements to steer electromagnetic waves, IRS can reconfigure wireless environments to enhance coverage without the substantial cost and power overhead of active RF chains. However, most theoretical research assumes ideal hardware. In practice, Access Points (APs) suffer from RF impairments—such as I/Q imbalance and amplifier non-linearities—while IRS elements exhibit phase noise due to the use of low-resolution, discrete phase shifters.

This paper investigates how these dual hardware imperfections impact the performance of IRS-assisted Multiple-Input Single-Output (MISO) downlink systems. The study reveals a critical "hardware ceiling": unlike ideal systems where spectral efficiency (SE) grows indefinitely with more hardware, non-ideal systems hit a hard upper bound.

Methodology

The researchers employed a rigorous analytical framework to model the non-ideal system:

  • AP Impairment Modeling: An Extended Error Vector Magnitude (EEVM) model was used to capture aggregate RF impairments, including attenuation, phase rotation, and additive distortion noise.
  • IRS Impairment Modeling: Phase noise at the IRS was modeled as a uniformly distributed random error, representing the limitations of practical, low-resolution phase shifters.
  • Beamforming and Analysis: The system assumes Maximum Ratio Transmission (MRT) at the AP. IRS phase shifts are optimized based on ideal hardware assumptions, reflecting practical scenarios where instantaneous errors are unknown. Asymptotic analysis using the Strong Law of Large Numbers was performed to derive closed-form expressions for SE as the number of antennas ($M$) and reflecting elements ($N$) approach infinity.
  • Energy Efficiency (EE): The authors formulated an EE maximization problem, solving for the optimal transmit power using the Lambert W-function.

Spectral Efficiency and the Hardware Ceiling

A primary finding is that hardware impairments impose a fundamental limit on spectral efficiency. In an ideal system, increasing the number of reflecting elements $N$ leads to a logarithmic increase in SE. However, in the presence of impairments, SE converges to a finite value.

Spectral efficiency versus SNR (top) and number of reflecting elements N (bottom), illustrating the performance ceiling caused by hardware impairments.

Interestingly, at high Signal-to-Noise Ratios (SNR), the performance bottleneck is primarily dictated by the AP's RF impairments. The impact of IRS phase noise mathematically diminishes in this regime. This suggests that network operators can deploy cost-effective, low-resolution IRS panels without significantly compromising high-SNR performance, as the IRS beamforming effectively cancels out some of its own phase noise penalties relative to the dominant AP distortion.

Energy Efficiency Optimization

The study also explores the trade-offs in energy efficiency. As hardware impairments increase, the peak achievable energy efficiency decreases. Furthermore, the system requires a higher optimal transmit power ($P^*$) to reach its (lower) maximum EE compared to an ideal system.

Energy efficiency versus transmit power for various impairment levels, showing the shift in optimal transmit power P*.

The authors derived the exact, unique optimal transmit power $P^*$ in closed form using the Lambert W-function, providing a practical tool for system design in non-ideal conditions.

Conclusion

This research highlights that hardware impairments fundamentally change the scaling laws of IRS-assisted networks. While adding more antennas or reflecting elements cannot overcome the impairment-induced spectral efficiency ceiling, the relative insignificance of IRS phase noise at high SNR provides a strong economic justification for using lower-cost IRS components.

Future work may involve extending this analysis to multi-user MIMO setups and more complex fading environments, such as Rician or Rayleigh channels, to better capture non-Line-of-Sight scenarios.

Summary updated

Abstract

In this letter, we analyze the spectral and energy efficiency of an intelligent reflecting surface (IRS)-assisted multiple-input single-output (MISO) downlink system with hardware impairments. An extended error vector magnitude (EEVM) model is utilized to characterize the impact of radio-frequency (RF) impairments at the access point (AP) and phase noise is considered at the IRS. We show that the spectral efficiency is limited due to the hardware impairments even when the numbers of AP antennas and IRS elements grow infinitely large, which is in contrast with the conventional case with ideal hardware. Moreover, the performance degradation at high SNR is shown to be mainly affected by the AP hardware impairments rather than by the phase noise at the IRS. We further obtain in closed form the optimal transmit power for energy efficiency maximization. Simulation results are provided to verify the obtained results.
BibTeX
@article{zhou2020spectral,
  author = {Shaoqing Zhou and Wei Xu and Kezhi Wang and Marco Di Renzo and Mohamed-Slim Alouini},
  title = {{Spectral and Energy Efficiency of IRS-Assisted MISO Communication With Hardware Impairments}},
  journal = {IEEE Wireless Communications Letters},
  year = {2020},
  url = {https://openreview.net/forum?id=ok9tAwyy3y5},
}
Figure