Performance Evaluation of Beyond Diagonal RIS under Hardware Impairments
Jose Carlos da Silva Filho, Josué V. de Araújo, Bruno Sokal, André L. F. de Almeida · 2025
Read Performance Evaluation of Beyond Diagonal RIS under Hardware Impairments by Jose Carlos da Silva Filho, Josué V. de Araújo, Bruno Sokal, André L. F. de Almeida (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2506-07266.txt.
Introduction
Jose Carlos da Silva Filho and co-authors present a 2025 engineering paper on smart radio surfaces. Reconfigurable intelligent surfaces steer wireless signals by tuning many small elements. A beyond-diagonal design links the elements together for finer wave control. Real hardware is imperfect: impedances mismatch and tuning diodes misbehave. The authors build three models of such flaws. Each flaw type hits the surface's scattering behavior in its own way.
Channel estimation error is the scoreboard for the damage. Some flaw patterns hurt the advanced design more than they hurt the plain one. In those cases the simpler traditional surface actually wins. Wireless engineers get a caution with structure: added control pays only when hardware quality supports it. Deployment budgets can then split between smarter surfaces and cleaner components with open eyes. The impairment models are reusable for other surface designs.
Catalog record
| CREATOR | Jose Carlos da Silva Filho, Josué V. de Araújo, Bruno Sokal, André L. F. de Almeida |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 26 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2506-07266
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Performance Evaluation of Beyond Diagonal RIS under Hardware Impairments Jose Carlos da Silva Filho, Josué V. de Araújo, Bruno Sokal, André L. F. de Almeida Abstract Beyond diagonal reconfigurable intelligent surface (BD-RIS) improves the traditional reconfigurable intelligent surface (RIS) architecture functionality by interconnecting elements for advanced wave control. However, real-world implementations face hardware imperfections, such as impedance mismatches and varactor nonidealities, which can degrade overall system performance. In this paper, we propose three hardware impairment models that directly affect the BD-RIS scattering matrix structure and evaluate their impact on the channel estimation accuracy using the normalized mean square error (NMSE) as a performance metric. The proposed impairment models consider imperfections affecting self-impedances, mutual impedances, or both. Our results reveal how each impairment type degrades the system performance, allowing us to identify scenarios where the traditional RIS can outperform the BD-RIS. keywords Beyond diagonal RIS, hardware impairments, channel estimation. 1 Introduction In a typical communication system, the transmitted signal suffers from several impairments due to the surrounding environment, including attenuation and scattering, which is generally caused by multipath propagation, one of the main limiting factors in this wireless communication system. An emerging technology capable of mitigating these effects and improving propagation conditions is RIS which is a man-made surface composed of electromagnetic materials that consists of a 2D array with many low-cost passive elements that can control the electromagnetic properties of radio-frequency waves, such as phase, amplitude, and frequency [ 1 , 2 , 3 ] . In general, the propagation of electromagnetic waves in wireless communication channels is not controlled between the transmission and reception processes. However, the RIS can "reprogram" the channel during this stage. Classical RIS architectures consider each element independent of the others, i.e., no interconnection among elements exists. This design translates into the signal model, where the scattering matrix has a diagonal structure. As a result, some limitations arise, such as limited beamforming capabilities and restricted degrees of freedom to control the direction of signal reflection [ 4 ] . Recent studies have introduced beyond diagonal reconfigurable intelligent surface (BD-RIS) to overcome conventional RIS’s limitations and enhance system performance [ 5 ] . In BD-RIS architectures, the reflecting elements are intentionally interconnected, leading to a non-diagonal signal model matrix. More precisely, in the BD-RIS architecture, the connections are primarily defined as [ 6 , 7 , 8 , 9 , 10 , 11 ] 1) group-connected, where only a portion of the RIS elements are interconnected, leading to the signal model a block-diagonal matrix as scattering matrix, and 2) fully-connected, which represents the design that all RIS elements are connected. In this case, the signal model scattering matrix is complete, i.e., with all non-zero entries. This non-diagonal design of the scattering matrix provides additional degrees of freedom and enables a more flexible control over the signal properties. This includes joint manipulation of phase and amplitude and the ability to create more reflection patterns, which, in specific scenarios, allow BD-RIS to outperform the conventional diagonal RIS. Despite the promising theoretical performance of BD-RIS, most existing studies are still based on idealized models with perfect operating conditions. In many cases, however, such models do not fully capture the practical imperfections possible in real-world implementations, such as impedance mismatches and losses in the passive interconnected components of the internal circuits of the BD-RIS structure. These limitations highlight the need for more realistic performance evaluations considering physical/electronic impairments. In this paper, we evaluate the performance of BD-RIS-assisted communication systems in the presence of circuit-based hardware impairments. We start from the baseline signal model of [ 9 ] by including the hardware impairment models and assess the solution of [ 9 ] under imperfections. We consider three hardware impairment models directly affecting the self-impedances, mutual impedances, or both. Simulation results show how these imperfections affect the channel estimation performance when state-of-the-art algorithms are used. 2 Notation and properties To facilitate the understanding of the material presented throughout the paper, we define the key mathematical notations and properties used. Scalars are denoted by non-bold lowercase letters, such as a , column vectors are denoted by bold lowercase letters, such as \boldsymbol{a} , and matrices are represented by bold uppercase letters, such as \boldsymbol{A} . \boldsymbol{I}_{K} represents an identity matrix of size K\times K , while \boldsymbol{1}_{N\times N} stands for a all-ones matrix of size N\times N . The superscripts \{\cdot\}^{\mathrm{T}},\{\cdot\}^{*} and \{\cdot\}^{\mathrm{H}} denote the transpose, conjugate, and conjugate transpose, respectively. The operator \|\cdot\|_{F} refers to the Frobenius norm of a matrix or tensor, and \mathbb{E}\{\cdot\} denotes the expectation operator. The operator \operatorname{diag}(\boldsymbol{a}) converts a vector \boldsymbol{a} into a diagonal matrix. From a set of Q matrices \boldsymbol{X}^{(q)}\in\mathbb{C}^{M\times N} , with q=\{1,\ldots,Q\} , we can construct a block diagonal matrix as \boldsymbol{X}=\operatorname{blkdiag}\left(\boldsymbol{X}^{(1)},\ldots,\boldsymbol{X}^{(Q)}\right)\in\mathbb{C}^{MQ\times NQ} . Moreover, the operator \operatorname{vec}(\boldsymbol{A}) converts a matrix \boldsymbol{A}\in\mathbb{C}^{M\times N} to a column vector \boldsymbol{a}\in\mathbb{C}^{MN\times 1} by stacking its columns on top of each other, while the operator unvec (\boldsymbol{a})_{M\times N} reverses this operation, restoring the matrix \boldsymbol{A}\in\mathbb{C}^{M\times N} . The symbols \circ , \odot , \otimes , and \diamond denote the outer product, the Hadamard product, the Kronecker product, and the Khatri-Rao product (also known as the column-wise Kronecker product), respectively. Some useful properties that will be exploited in this paper are the following. \displaystyle\operatorname{vec}(\boldsymbol{A}\boldsymbol{B}\boldsymbol{C}) \displaystyle=\left(\boldsymbol{C}^{\mathrm{T}}\otimes\boldsymbol{A}\right)\operatorname{vec}(\boldsymbol{B}), (1) \displaystyle\operatorname{vec}(\boldsymbol{A}\otimes\boldsymbol{B}) \displaystyle=\boldsymbol{P}(\operatorname{vec}(\boldsymbol{A})\otimes\operatorname{vec}(\boldsymbol{B})), (2) \displaystyle\operatorname{vec}\left(\boldsymbol{a}\boldsymbol{b}^{\mathrm{T}}\right) \displaystyle=\boldsymbol{b}\otimes\boldsymbol{a}, (3) where the vectors and matrices involved have compatible dimensions in each case, and \boldsymbol{P} is a permutation matrix. The Khatri-Rao product of the matrices \boldsymbol{X}\in\mathbb{C}^{I\times R} and \boldsymbol{Y}\in\mathbb{C}^{J\times R} , is defined as \displaystyle\boldsymbol{Z}=\boldsymbol{X}\diamond\boldsymbol{Y}=\left[\boldsymbol{x}_{1}\otimes\boldsymbol{y}_{1},\ldots,\boldsymbol{x}_{R}\otimes\boldsymbol{y}_{R}\right]\in\mathbb{C}^{JI\times R}, (4) where \boldsymbol{x}_{r} and \boldsymbol{y}_{r} , are the r -th column of \boldsymbol{X} and \boldsymbol{Y} , respectively, for r=\{1,\ldots,R\} . 3 System Model We consider the system model described in [ 12 ] , i.e., a multiple-input multiple-output (MIMO) communication system assisted by a BD-RIS composed of N elements, divided into Q groups, each of size \bar{N} , i.e., N=\bar{N}Q . The transmitter (Tx) and the receiver (Rx) have M_{T} and M_{R} antennas, respectively. The direct link (TX-RX) is assumed to be blocked for convenience. The transmission duration is over T time-slots, from which the transmitted pilots are defined as \boldsymbol{X}=[\boldsymbol{x}_{1},\ldots,\boldsymbol{x}_{T}]\in\mathbb{C}^{M_{T}\times T} . The received signal on the t -th transmitted time slot is given by \displaystyle\boldsymbol{y}_{t}=\sum\limits_{q=1}^{Q}\boldsymbol{G}^{(q)}\boldsymbol{S}_{t}^{(q)}\boldsymbol{H}^{(q)\text{T}}\boldsymbol{x}_{t}+\boldsymbol{b}_{t}\in\mathbb{C}^{M_{R}\times 1}, (5) where, \boldsymbol{H}^{(q)}\in\mathbb{C}^{M_{T}\times\bar{N}} and \boldsymbol{G}^{(q)}\in\mathbb{C}^{M_{R}\times\bar{N}} are the q -th block matrix of the TX-RIS and RIS-RX channels, i.e., \boldsymbol{H}=[\boldsymbol{H}^{(1)},\ldots,\boldsymbol{H}^{(Q)}]\in\mathbb{C}^{M_{T}\times\bar{N}Q} and \boldsymbol{G}=[\boldsymbol{G}^{(1)},\ldots,\boldsymbol{G}^{(Q)}]\in\mathbb{C}^{M_{R}\times\bar{N}Q} . \boldsymbol{S}_{t}^{(q)}\in\mathbb{C}^{\bar{N}\times\bar{N}} represents the BD-RIS scattering matrix of the q -th group at the t -th time slot. To model a possible real-world hardware imperfection affecting the impedances, we consider different types of impairments that are given by a matrix \mathbf{E}\in\mathbb{C}^{N\times N} in which is detailed in Section 4 . For the sake of convenience, we consider that hardware impairment is constant over the transmission of the T time slots. Thus, the real observed BD-RIS scattering matrix, at the t -th time slot, can be denoted as \displaystyle\bar{\boldsymbol{S}}_{t}=\boldsymbol{S}_{t}\odot\boldsymbol{E}\in\mathbb{C}^{N\times N}, (6) where \boldsymbol{S}_{t} is the ideal BD-RIS scattering matrix at the t -th time slot, for t=1,\ldots,T . Hence, the signal received in the t -th time slot can be expressed as \displaystyle\boldsymbol{y}_{t} \displaystyle=\sum\limits_{q=1}^{Q}\boldsymbol{G}^{(q)}\left(\boldsymbol{S}_{t}^{(q)}\odot\boldsymbol{E}\right)\boldsymbol{H}^{(q)\text{T}}\boldsymbol{x}_{t}+\boldsymbol{b}_{t}\in\mathbb{C}^{M_{R}\times 1}, (7) \displaystyle=\sum\limits_{q=1}^{Q}\boldsymbol{G}^{(q)}\bar{\boldsymbol{S}}_{t}\boldsymbol{H}^{(q)\text{T}}\boldsymbol{x}_{t}+\boldsymbol{b}_{t}. (8) Figure 1: System model of a MIMO communication aided by a BD-RIS. Figure 2: Illustration