Performance Analysis of Wireless Communication Systems Assisted by Fluid Reconfigurable Intelligent Surfaces

Chan-Byoung Chae; Farshad Rostami Ghadi; F. Javier Lopez-Martinez; George C. Alexandropoulos; Kai-Kit Wong

REVIEW 4 major objections 6 minor 1 cited by

Performance Analysis of Wireless Communication Systems Assisted by Fluid Reconfigurable Intelligent Surfaces

T0 review · 4 major / 6 minor · reviewed 2026-08-07 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

Pith's one-line read A fluid RIS that dynamically selects its reflecting elements can substantially improve outage probability and ergodic capacity over a fixed RIS, and the paper provides closed-form Gamma approximations for both.

desk verdict The Gamma approximation for the FRIS channel is built on a false trace identity, so the paper's central performance claims are unsupported. read the letter →

arxiv 2505.23680 v1 pith:7DSS7LCF submitted 2025-05-29 cs.IT eess.SPmath.IT

Farshad Rostami Ghadi , Kai-Kit Wong , F. Javier Lopez-Martinez , George C. Alexandropoulos , Chan-Byoung Chae This is my paper

classification cs.ITeess.SPmath.IT

keywords fluidreconfigurableintelligentsurfaceFRISoutageprobabilityergodiccapacitymomentmatchingGammadistributionspatialcorrelationelementselection

verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

The paper claims that a fluid reconfigurable intelligent surface (FRIS), a RIS whose densely packed elements can be dynamically switched on or off to select the best subset for reflection, can substantially improve link reliability and spectral efficiency compared with a conventional fixed-element RIS. To make this tractable, it approximates the end-to-end channel power gain by a Gamma distribution using moment matching, with shape and scale parameters expressed through traces of the spatial correlation submatrix of the selected elements. From this approximation it derives closed-form expressions for outage probability and an upper bound on ergodic capacity, plus their high-SNR asymptotics. If the approximation holds, FRIS becomes an analytically tractable design whose gains come from exploiting spatial diversity through element selection rather than from adding more fixed elements.

What carries the argument

The engine of the proof is moment matching of the quadratic form $G=h_f^H B h_f$, where $B=A^H h_u h_u^H A$ and $A=J^{1/2}S_{M_o}^T\Phi S_{M_o}J^{1/2}$. Lemma 1 identifies $G$ as generalized chi-squared because it is a Hermitian quadratic form in a complex Gaussian vector; Proposition 1 retains only the first two moments, $\mathbb{E}[G]=\operatorname{tr}(AA^H)$ and $\mathbb{E}[G^2]=\operatorname{tr}^2(AA^H)+\operatorname{tr}((AA^H)^2)$, and matches them to a Gamma density. The traces are then reduced using the identities $\operatorname{tr}(\tilde\Phi\tilde J\tilde\Phi^H\tilde J)=\operatorname{tr}(\tilde J^2)$ and $\operatorname{tr}((\tilde\Phi\tilde J\tilde\Phi^H)^2\tilde J^2)=\operatorname{tr}(\tilde J^4)$, asserted because $\tilde\Phi$ is diagonal with unit-modulus entries. All later metrics therefore depend only on the reduced spatial correlation matrix $\tilde J$ of the selected elements.

What would settle it

For a two-element setup with $\tilde J=[[1,\rho],[\rho,1]]$ and phases $(1,-1)$, compute $\operatorname{tr}(\tilde\Phi\tilde J\tilde\Phi^H\tilde J)=2(1-\rho^2)$, which differs from $\operatorname{tr}(\tilde J^2)=2(1+\rho^2)$. A Monte-Carlo outage simulation for that configuration would therefore disagree with formula (20), settling whether the trace identity holds for general phase selections.

Watch

Extended reading notes

Core claim

At the center is a Gamma approximation for the equivalent channel power gain $G=|h_u^H J^{1/2} S_{M_o}^T \Phi S_{M_o} J^{1/2} h_f|^2$, a quadratic form in complex Gaussian vectors. Moment matching gives $G\sim\mathrm{Gamma}(k,\theta)$ with $k=(\operatorname{tr}(\tilde J^2))^2/\operatorname{tr}(\tilde J^4)$ and $\theta=\operatorname{tr}(\tilde J^4)/\operatorname{tr}(\tilde J^2)$, where $\tilde J=S_{M_o} J S_{M_o}^T$ is the correlation submatrix of the activated elements. From this the paper derives the outage probability $P_o=\frac{1}{\Gamma(k)}\Upsilon(k,\bar R/(\gamma L_f L_u\theta))$ and the ergodic capacity upper bound $\log_2(1+\gamma L_f L_u\operatorname{tr}(\tilde J^2))$. The numerical comparisons in the paper show FRIS outperforming conventional RIS at equal element counts, with ergodic capacity about 34% higher at $\gamma=40$ dB for $M_o=16$ and outage near $10^{-4}$ versus above $10^{-3}$ at $M_o=36$.

Load-bearing premise

The outage and capacity formulas assume that the chosen reflection phases do not alter two particular aggregate sums of the correlation matrix of the activated elements. This assumption fails for some phase patterns, and when it fails the closed-form expressions no longer describe the channel.

Editorial extensions

If this is right

Outage probability and ergodic capacity of FRIS-assisted links become closed-form functions of the activated-element correlation submatrix, so system-level evaluation no longer requires Monte-Carlo simulation.
Because the capacity bound depends only on $\operatorname{tr}(\tilde J^2)$, the choice of which elements to activate can be cast as a trace-maximization problem over the aperture's correlation structure.
Raising the total surface density $M$ while keeping the number of active elements fixed still grows capacity, whereas a conventional RIS with fixed elements gains nothing from a denser grid without more elements.
The numerically observed saturation as $M_o$ grows implies that the FRIS advantage is largest at moderate active-element counts and in the high-SNR regime, where diversity is already rich.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

The trace-only form of the capacity bound suggests a practical activation rule that depends only on the geometry of the surface, not on instantaneous channel realizations: activate the subset with the largest $\operatorname{tr}(\tilde J^2)$ and align phases afterward.
Since $G$ is a generalized chi-square, the same two-moment Gamma matching should extend to other quadratic-form performance metrics in fluid-antenna and intelligent-surface systems, such as secrecy outage, with the same trace machinery.
The trace identities behind the Gamma parameters are not true for all phase patterns, so an extension that adds phase-dependent correction terms would be needed to keep the formulas accurate under arbitrary discrete phase optimization.

Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Desk Editor's Note

read the letter

Quick read of arXiv:2505.23680. The paper applies a standard Gamma moment-matching approximation to an FRIS channel and derives outage and capacity formulas; that is a reasonable extension, and the Monte Carlo effort is visible. But there is a load-bearing error that collapses the analysis. The step from (16) to (17) asserts tr(Φ̃J̃Φ̃^H J̃) = tr(J̃^2) for any unit-modulus diagonal Φ̃. That is false. For J̃ = [[1,ρ],[ρ,1]] and Φ̃ = diag(1,−1), the left side is 2(1−ρ²) while the right side is 2(1+ρ²). The same phase dependence carries into (19) and therefore into the Gamma shape and scale, the outage probability, and the capacity bound. The agreement with simulation in Fig. 2 does not fix this; it just means the simulations and theory share the same erroneous trace evaluation.

There is also a modeling inconsistency. The paper describes the selection matrix as choosing elements to maximize the equivalent gain, but the statistical derivation treats the selection matrix as fixed and channel-independent. So even if the trace identity were corrected, the derivation would not describe the optimized system whose gains are reported. The claimed FRIS advantage is essentially inserted through the assumption that selection is perfect, without modeling that selection in the statistics.

What is genuinely useful: the FRIS channel model with spatial correlation, the observation that G is a generalized chi-squared quadratic form, and the asymptotic outage expressions are sensible. The citation pattern looks fine, and the paper does not hide the approximations. But the central moment-matching derivation fails on its own terms, so the quantitative results should not be used.

For peer review, I would send it to a referee to confirm the trace error and, if the authors can correct it, allow a major revision. As written, I would not cite it. For a reading group, I would skip it unless you want a concrete example of a trace identity pitfall.

Recommendation: desk reject with a clear explanation, or send to a referee for a second check if the editor wants to be thorough.

Referee Report

4 major / 6 minor

Summary. The manuscript analyzes a downlink wireless link assisted by a fluid reconfigurable intelligent surface (FRIS). It models the equivalent channel gain G = |h_u^H J^{1/2} S_{M_o}^T \Phi S_{M_o} J^{1/2} h_f|^2, approximates the distribution of G by a Gamma distribution using moment matching, and then derives closed-form approximations for outage probability and ergodic capacity, including high-SNR asymptotics. Numerical simulations are presented to support the claim that FRIS substantially improves reliability and spectral efficiency over conventional RIS.

Significance. If the derivations were correct, the paper would provide a useful closed-form statistical characterization for FRIS-assisted links, a topic of current interest. The paper includes Monte Carlo validation and clear system modeling, which are strengths. However, the central moment derivations contain load-bearing algebraic errors: the second moment of G is computed incorrectly, and the trace simplifications that eliminate the phase dependence are false for general unit-modulus phase shifts. These errors invalidate the Gamma shape and scale parameters, the outage probability, the capacity bound, and the numerical comparisons that rely on them. As a result, the main performance claims are not supported by the analysis as written.

major comments (4)

[Section III-A, Eqs. (16)-(17)] The equality tr(AA^H) = tr(\tilde J^2) is false for general unit-modulus phase matrices. For a real symmetric 2x2 matrix \tilde J = [[1, rho],[rho,1]] and D = diag(1,-1), one obtains tr(D \tilde J D^H \tilde J) = 2(1-rho^2), whereas tr(\tilde J^2) = 2(1+rho^2). In general, tr(D \tilde J D^H \tilde J) = sum_i \tilde J_{ii}^2 + 2 sum_{i<j} Re(D_i \bar D_j) \tilde J_{ij}^2, which equals tr(\tilde J^2) only when Re(D_i \bar D_j)=1 for all correlated pairs, i.e., when all selected phase shifts are equal modulo 2pi. Since the manuscript explicitly assumes phase shifts are optimized to coherently align reflected signals, the phases are generally not equal, so the simplification is invalid in the regime of interest. The same phase dependence invalidates Eq. (19): for the same example, tr((D \tilde J D^H \tilde J)^2) = 2(1-rho^2)^2, whereas tr(\tilde J^4) = (1+rho)^4+(1-rho)^4. Consequently, the Gamma parameters k and theta in Proposition 1, the outage probability in (20), and the capacity bound in (24) are unsupported.
[Section III-A, Eq. (12)] The second moment E[G^2] is computed incorrectly. For the scalar case M=1 with A=a, the true value is E[G^2] = E[|h_u|^4] E[|h_f|^4] |a|^4 = 4|a|^4, whereas Eq. (12) gives 2|a|^4. More generally, condition on h_u: given h_u, the random variable h_u^H A h_f is complex Gaussian with variance h_u^H A A^H h_u, so |h_u^H A h_f|^2 is exponential with that mean. Therefore E[G^2|h_u] = 2 (h_u^H A A^H h_u)^2, and taking the expectation over h_u yields E[G^2] = 2 tr^2(AA^H) + 2 tr((AA^H)^2). This missing factor of two changes the variance, and hence the Gamma shape parameter k in Proposition 1, independently of the trace-identity issue raised above.
[Section II and Section III-A] The system model states that the selection matrix S_{M_o} models the dynamic activation of the M_o elements that maximize the equivalent channel gain, but the analytical derivation treats S_{M_o} as a deterministic projection that is independent of the fading channels. The trace expressions in Proposition 1 depend only on the principal submatrix \tilde J, with no order-statistic or optimization term appearing. Thus the derived formulas do not capture the selection gain that is credited to FRIS in the numerical comparisons of Section IV. To substantiate the claimed gains, the analysis would need to model the distribution of \tilde J under optimal (or at least random) element selection, rather than treating the selection as fixed.
[Section III-A, Lemma 1] The proof of Lemma 1 defines B = A^H h_u h_u^H A and then treats B as a fixed positive semidefinite matrix when concluding that G follows a generalized chi-squared distribution with deterministic weights zeta_l. However, B depends on the random vector h_u, so the unconditional distribution of G is not a weighted sum of independent exponentials with fixed coefficients. The statement is at best valid conditionally on h_u. This does not directly affect the moment computations in Proposition 1, but it is an error in a numbered lemma and should be corrected.

minor comments (6)

[Section II, Eq. (1) and footnote 1] The indexing in Eq. (1) is inconsistent with the footnote: the main text uses mod(i, M_x) and floor(i/M_x), while the footnote states that row-major indexing uses mod(i-1, M_x) and floor((i-1)/M_x). Please align the notation throughout.
[Section II, after Eq. (2)] J_0 is the zero-order Bessel function of the first kind, not the zero-order spherical Bessel function; please correct the terminology.
[Section II, system model] In the description of the FRIS surface size, the total length of each column is written as W_x\lambda, which should presumably be W_z\lambda.
[Section IV] The numerical section does not specify how the M_o active elements are selected in the Monte Carlo simulations (e.g., exhaustive search, greedy algorithm, or random selection). This information is needed to reproduce the results and to verify that the simulations correspond to the optimized selection described in the system model.
[Section III-C, Proposition 3] The bound in (24) is described as 'tight' without a quantitative statement; a figure-based observation is not a proof of tightness. Please either remove the word 'tight' or provide a bound on the gap.
[Funding footnote and references] There is a typo 'supported by by' in the funding footnote, and reference [10] lists an unusual arXiv identifier 'arXiv:6455283' that should be checked.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity found: the Gamma approximation is derived from assumed model moments rather than fitted to the data being predicted, and the later outage and capacity expressions are algebraic consequences of that approximation. The suspicious trace identities are a mathematical-error concern, not a circularity.

full rationale

The central derivation chain is Proposition 1: starting from the system model G = |h_u^H A h_f|^2, the paper computes the first two moments using Gaussian Wick/Isserlis identities, matches those moments to a Gamma distribution, and then substitutes the resulting k and theta into the outage CDF (Eq. 20) and capacity bound (Eq. 24). No parameter is fitted to outage or capacity data: k and theta are deterministic functions of the assumed correlation matrix and selection matrix, and the Monte-Carlo comparisons are external validation rather than inputs. The cited prior work (Wong et al. on FAS; the authors' own RIS-FAS papers) is used only as background or as the source of the FRIS concept, not as justification for the moment-matching result; in particular, no uniqueness theorem or load-bearing result is imported from the authors' prior papers. The only serious defect identified by a careful reading is the algebraic step between Eqs. (16)-(17) and (18)-(19), where tr(tilde Phi tilde J tilde Phi^H tilde J) is asserted equal to tr(tilde J^2) for arbitrary unit-modulus diagonal tilde Phi; as the counterexample with tilde J = [[1,rho],[rho,1]] and tilde Phi = diag(1,-1) shows, this identity is false in general. That is a correctness or validity flaw in the derivation, but it is not circularity: the formulas do not reduce to their own inputs, nor is a fitted parameter renamed as a prediction, nor is the conclusion forced by self-citation. Accordingly, under the circularity rubric that distinguishes mathematical error from circular reasoning, the appropriate score is 0.

Assumptions & free parameters 0 free parameters · 6 assumptions · 0 invented entities

The paper does not fit any numerical parameters to data; the Gamma shape and scale are derived from assumed model moments. It relies on several domain assumptions (Jakes correlation, i.i.d. fading, blocked direct link) and on two ad hoc assumptions central to the derivation: the trace invariance under phase rotations, which is false in general, and the claim that a fixed-subset quadratic form approximates an optimized-selection system. No new physical entities are introduced.

assumptions (6)

domain assumption Spatial correlation between FRIS elements follows Jakes' model j_{i,j}=J0(2*pi*d_{i,j}/lambda).
Introduced in Section II after Eq. (1); it defines the correlation matrix J on which all trace computations and the Gamma parameters depend.
domain assumption Small-scale fading vectors h_f and h_u are i.i.d. CN(0,I_M) and independent of each other.
Stated in Section II; this makes the equivalent channel a quadratic form in Gaussian variables and enables the Isserlis theorem step in Proposition 1.
domain assumption The direct BS-MU link is blocked.
Stated in Section II; removal of the direct link defines the relayed-only SNR model and excludes a more general comparison.
ad hoc to paper Phase shifts are ideal and fully controllable, and the selection matrix picks the Mo elements that maximize the instantaneous equivalent channel gain.
This is the defining FRIS mechanism per Section II, but Section III treats S and Phi as deterministic and channel-independent, so the optimization is not actually reflected in the derived statistics.
ad hoc to paper The trace identities tr(AA^H)=tr(tilde J^2) and tr((AA^H)^2)=tr(tilde J^4) hold for arbitrary unit-modulus phase rotations.
Used in equations (17) and (19). This is false in general; a 2x2 correlated submatrix with phases (1,-1) gives a different trace, so the Gamma parameters are not justified.
ad hoc to paper The equivalent channel gain is well approximated by a Gamma distribution with matched first two moments.
Moment-matching is an approximation and no error bound is provided. Figure 2 validates one fixed configuration, not the optimized-selection distribution, so the approximation's accuracy for the claimed system is unproven.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Performance Analysis of Wireless Communication Systems Assisted by Fluid Reconfigurable Intelligent Surfaces." pith.science (2026). https://pith.science/paper/7DSS7LCF

@misc{pith2026250523680,
  author       = {Pith},
  title        = {Pith review of: Performance Analysis of Wireless Communication Systems Assisted by Fluid Reconfigurable Intelligent Surfaces},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/7DSS7LCF}},
  note         = {Machine review of arXiv:2505.23680}
}

read the original abstract

This letter investigates the performance of emerging wireless communication systems assisted by a fluid reconfigurable intelligent surface (FRIS). Unlike conventional reconfigurable intelligent surfaces (RISs), an FRIS consists of fluid-inspired metamaterials arranged in a densely packed matrix of sub-elements over a surface. It dynamically activates specific elements for signal reflection and modulation based on real-time channel conditions. Considering a downlink scenario where a base station communicates with a user terminal via a FRIS, we first characterize the statistical behavior of the equivalent end-to-end channel by deriving closed-form approximations for its cumulative distribution and probability density functions. Using these expressions, an analytical approximation for the outage probability and a tight upper bound on the ergodic capacity, including their asymptotic behaviors for high signal-to-noise ratio values, are derived. Our findings reveal key performance trends demonstrating that FRIS can substantially improve link reliability and spectral efficiency compared to conventional RISs, owing to its capability to dynamically select optimal elements from a dense preconfigured grid.

Figures

Figures reproduced from arXiv: 2505.23680 by the authors.

**Figure 2.** Illustration of (a) the PDF and (b) the CDF of the equivalent channel [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗

**Figure 3.** (a) Outage probability performance versus the average SNR [PITH_FULL_IMAGE:figures/full_fig_p005_3.png] view at source ↗

Discussion (0). Sign in to comment.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

Fluid Antenna Systems: A Geometric Approach to Error Probability and Fundamental Limits
cs.IT 2025-09 reject novelty 5.0 of 10

Derives an asymptotic SER formula for fluid antennas and claims diversity gain is set by effective rank ≈ 2W+1 (aperture width), not port count.

Reference graph

Works this paper leans on

12 extracted references · 9 canonical work pages · cited by 1 Pith paper

[1]

Reconfigurable intelligent surfaces for 6G: Emerging applications and open challenges,

E. Basar et al. , “Reconfigurable intelligent surfaces for 6G: Emerging applications and open challenges,” IEEE Veh. Technol. Mag. , vol. 19, no. 3, pp. 27–47, 2024

work page 2024
[2]

Reconfigurable intelligent surfaces for rich scattering wireless communications: Recent experiments, chal- lenges, and opportunities,

G. C. Alexandropoulos et al. , “Reconfigurable intelligent surfaces for rich scattering wireless communications: Recent experiments, chal- lenges, and opportunities,” IEEE Commun. Mag. , vol. 59, no. 6, pp. 28–34, 2021

work page 2021
[3]

Reconfigurable intelligent surfaces for wireless com- munications: Overview of hardware designs, channel models, and esti- mation techniques,

M. Jian et al. , “Reconfigurable intelligent surfaces for wireless com- munications: Overview of hardware designs, channel models, and esti- mation techniques,” Intelligent & Conv. Netw. , vol. 3, no. 1, pp. 1–32, 2022

work page 2022
[4]

Fluid antenna systems,

K. K. Wong et al. , “Fluid antenna systems,” IEEE Trans. Wireless Commun., vol. 20, no. 3, pp. 1950–1962, Mar. 2021

work page 1950
[5]

A tutorial on fluid antenna system for 6G networks: Encompassing communication theory, optimization methods and hard- ware designs,

W. K. New et al., “A tutorial on fluid antenna system for 6G networks: Encompassing communication theory, optimization methods and hard- ware designs,” IEEE Commun. Surveys & Tuts. , doi:10.1109/COMST. 2024.3498855, 2024

arXiv 2024
[6]

On performance of RIS-aided fluid antenna systems,

F. Rostami Ghadi et al. , “On performance of RIS-aided fluid antenna systems,” IEEE Wireless Commun. Lett. , vol. 13, no. 8, pp. 2175–2179, Aug. 2024

2024
[7]

Secrecy performance analysis of RIS-aided fluid an- tenna systems,

F. R. Ghadi et al., “Secrecy performance analysis of RIS-aided fluid an- tenna systems,” in Proc. IEEE Wireless Commun. Netw. Conf. (WCNC) , Mar. 2025

work page 2025
[8]

A first look at the performance enhancement potential of fluid reconfigurable intelligent surface,

A. Salem et al., “A first look at the performance enhancement potential of fluid reconfigurable intelligent surface,” arXiv preprint, arXiv:2502. 17116v1, 2025

work page 2025

Show all 12 references

[9]

Fluid reconfigurable intelligent surfaces: Joint on-off selection and beamforming with discrete phase shifts,

H. Xiao et al. , “Fluid reconfigurable intelligent surfaces: Joint on-off selection and beamforming with discrete phase shifts,” arXiv preprint, arXiv:2503.14601v1, 2025

2025 arXiv
[10]

FIRES: Fluid integrated reflecting and emitting surfaces,

F. Rostami Ghadi et al., “FIRES: Fluid integrated reflecting and emitting surfaces,” arXiv preprint, arXiv:6455283, 2025

2025
[11]

Quadratic forms in random variables: Theory and applications,

A. M. Mathai and S. B. Provost, “Quadratic forms in random variables: Theory and applications,” J. American Statistical Assoc. , vol. 87, no. 420, Dec. 1992

1992
[12]

Gradshteyn and I

I. Gradshteyn and I. M. Ryzhik, Tables of Integrals, Series, and Products, Academic Press, 2000

2000

Pith tools

Reviewed August 7, 2026 · model on record in the stance chip above.

[1] [1]

Reconfigurable intelligent surfaces for 6G: Emerging applications and open challenges,

E. Basar et al. , “Reconfigurable intelligent surfaces for 6G: Emerging applications and open challenges,” IEEE Veh. Technol. Mag. , vol. 19, no. 3, pp. 27–47, 2024

work page 2024

[2] [2]

Reconfigurable intelligent surfaces for rich scattering wireless communications: Recent experiments, chal- lenges, and opportunities,

G. C. Alexandropoulos et al. , “Reconfigurable intelligent surfaces for rich scattering wireless communications: Recent experiments, chal- lenges, and opportunities,” IEEE Commun. Mag. , vol. 59, no. 6, pp. 28–34, 2021

work page 2021

[3] [3]

Reconfigurable intelligent surfaces for wireless com- munications: Overview of hardware designs, channel models, and esti- mation techniques,

M. Jian et al. , “Reconfigurable intelligent surfaces for wireless com- munications: Overview of hardware designs, channel models, and esti- mation techniques,” Intelligent & Conv. Netw. , vol. 3, no. 1, pp. 1–32, 2022

work page 2022

[4] [4]

Fluid antenna systems,

K. K. Wong et al. , “Fluid antenna systems,” IEEE Trans. Wireless Commun., vol. 20, no. 3, pp. 1950–1962, Mar. 2021

work page 1950

[5] [5]

A tutorial on fluid antenna system for 6G networks: Encompassing communication theory, optimization methods and hard- ware designs,

W. K. New et al., “A tutorial on fluid antenna system for 6G networks: Encompassing communication theory, optimization methods and hard- ware designs,” IEEE Commun. Surveys & Tuts. , doi:10.1109/COMST. 2024.3498855, 2024

arXiv 2024

[6] [6]

On performance of RIS-aided fluid antenna systems,

F. Rostami Ghadi et al. , “On performance of RIS-aided fluid antenna systems,” IEEE Wireless Commun. Lett. , vol. 13, no. 8, pp. 2175–2179, Aug. 2024

2024

[7] [7]

Secrecy performance analysis of RIS-aided fluid an- tenna systems,

F. R. Ghadi et al., “Secrecy performance analysis of RIS-aided fluid an- tenna systems,” in Proc. IEEE Wireless Commun. Netw. Conf. (WCNC) , Mar. 2025

work page 2025

[8] [8]

A first look at the performance enhancement potential of fluid reconfigurable intelligent surface,

A. Salem et al., “A first look at the performance enhancement potential of fluid reconfigurable intelligent surface,” arXiv preprint, arXiv:2502. 17116v1, 2025

work page 2025