Robust design for IRS-assisted multiuser systems under practical imperfections: a rate-splitting approach

Abstract

In practical intelligent reflecting surface (IRS)-assisted multiuser communication systems, inevitable imperfections such as hardware impairments, imperfect channel state information (CSI), and the limited resolution of the IRS phase shifts would introduce interference and thus cause significant performance degradation. As an interference management strategy, rate-splitting multiple access (RSMA) employs the rate-splitting (RS) principle to partition user information into common and private parts, thereby offering enhanced robustness. Accounting for practical imperfections, this study investigates robust beamforming design in IRS-assisted multiuser systems under the RSMA architecture. First, we introduce a system model that captures these non-ideal factors and evaluate their impacts on communication performance. To enhance the performance of the considered system, a weighted sum rate maximization problem is formulated, for which a sample average approximation (SAA)-based robust algorithm is proposed to jointly optimize the IRS phase shifts and the beamforming matrix at the base station (BS). Simulation results demonstrate that the IRS-assisted RSMA system exhibits superior robustness compared to the IRS-assisted space division multiple access (SDMA) system in the presence of inevitable imperfections. Furthermore, the proposed SAA-based robust algorithm outperforms existing benchmark algorithms, highlighting its effectiveness and robustness.

1. Introduction

The next-generation wireless communication systems aim to enhance existing 5G technologies while exploring novel solutions to support ultra-high data rates and massive connectivity (Tataria et al., 2021). Recently, intelligent reflecting surface (IRS) has emerged as a key technology for next-generation systems due to its low cost, wide coverage, and ability to adaptively control wireless propagation environments (Wu and Zhang, 2020; Wu et al., 2021; Zhang YP et al., 2023). An IRS is typically composed of a large array of passive reflecting elements, each capable of independently adjusting the amplitude and phase of the incident signal. Therefore, substantial passive gains can be provided without requiring additional power or active signal processing. Owing to these advantages, IRS can be seamlessly integrated with various advanced technologies to enhance system performance, such as massive multiple-input multiple-output (MIMO), integrated sensing and communication (ISAC), movable antenna (MA) technique, and orthogonal time frequency space (OTFS) modulation (Xu et al., 2021; Ma et al., 2024; Peng et al., 2024a, 2024b).

Despite its promising potential, the performance of IRS-assisted systems suffers significant degradation due to inevitable imperfections (Schenk, 2008; Björnson et al., 2013; Papazafeiropoulos et al., 2017; Kolawole et al., 2020; Wang ZR et al., 2020; Fu et al., 2021). First, the effectiveness of beamforming gain maximization is fundamentally limited by channel state information (CSI) acquisition accuracy. In IRS-assisted systems, CSI estimation is inherently challenging due to the limited number of pilot symbols and the passive nature of the IRS. These limitations lead to non-negligible CSI errors and introduce undesired interference, degrading overall system performance (Wang ZR et al., 2020). Moreover, inevitable hardware impairments in practical transceiver hardware can distort both transmitted and received signals (Papazafeiropoulos et al., 2017). While calibration and compensation techniques can partially mitigate these distortions, a certain amount of distortion and interference remains, significantly undermining the system's performance. Additionally, IRS phase shifts are inherently discrete rather than continuous, which makes it crucial to design efficient beamforming strategies that operate over discrete phase shift (DPS) sets (Fu et al., 2021). Overall, a robust algorithm design that explicitly accounts for these practical limitations is essential for exploiting the full potential of IRS-assisted communication systems.

Fortunately, rate-splitting multiple access (RSMA) technology demonstrates strong robustness against inevitable imperfections, enhancing the mutual benefits of combining IRS and RSMA (Aboumahmoud et al., 2025). Unlike space division multiple access (SDMA) and non-orthogonal multiple access (NOMA), where the former treats interference as noise and the latter relies on successive interference cancellation, RSMA splits user messages into common and private parts, enabling more flexible interference management. This paper focuses on the robust design of IRS-assisted multiuser systems under practical imperfections using an RSMA approach.