Identification of RIS-Assisted Paths for Wireless Integrated Sensing and Communication

Bashar Tahir; Markus Rupp; Stefan Schwarz; Zeyu Huang

REVIEW 3 major objections 5 minor 13 references

Identification of RIS-Assisted Paths for Wireless Integrated Sensing and Communication

T0 review · 3 major / 5 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 reconfigurable intelligent surface can tag its own reflection path by alternating part of its elements between coherent and random combining, letting a user identify the RIS path using only power measurements.

desk verdict Solid statistical model for detecting RIS pattern changes; the paper's central NLOS-identification claim is not actually modeled. read the letter →

arxiv 2506.04123 v1 pith:WAXRCEWG submitted 2025-06-04 eess.SP

Zeyu Huang , Stefan Schwarz , Bashar Tahir , Markus Rupp This is my paper

classification eess.SP

keywords reconfigurableintelligentsurfacesintegratedsensingandcommunicationpropagationpathidentificationdynamicRISconfigurationnon-centralchi-squareddistributionchannelpowerdetectionerrorprobabilityRIS-assisted

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 tries to establish that a reconfigurable intelligent surface (RIS) can encode a signature onto the reflection path it creates, so that a user device can tell signals that bounced off the RIS from signals bounced off ordinary scatterers. The scheme splits the RIS into a static part and a dynamic part; the dynamic part alternates between coherently focusing on the target user and on other users, which makes the estimated channel power of the RIS-assisted path swing between two known levels. The paper derives the exact statistical law of that power, a non-central chi-squared distribution with two degrees of freedom, and uses it to evaluate the detection error probability analytically. Simulation results confirm the analysis for different numbers of dynamic elements and different allocations of elements among users. If correct, this gives a low-complexity way to identify RIS paths for sensing and positioning without extra signaling overhead.

What carries the argument

The central object is the dynamic-part alternation of the RIS phase configuration: a subset of elements is switched between coherent combining for the target UE (pattern 1) and a configuration for other users, which from the target UE's viewpoint acts as random phases (pattern 2). The remaining elements are split into a static coherent area for the target UE and a static area configured for other users. The statistical machinery is the reduction, via the central limit theorem, of the random-phase sums to complex Gaussian random variables, so that the squared magnitude of the estimated channel becomes a non-central chi-squared random variable with two degrees of freedom. The distribution parameters are computed from the RIS-element channel amplitudes and the noise power, and the detection error probability follows by comparing the two CDFs with a threshold set between the two means.

What would settle it

In a full-wave simulation or testbed where the RIS element pathlosses vary strongly across the aperture (e.g., near-field or large element spacings), compute the empirical CDF of the normalized estimated channel power for both patterns and compare it to the non-central chi-squared fit from the paper's equations (19)-(20); a significant deviation would show that the CLT model, and the error probabilities derived from it, fail in that regime.

Watch

Extended reading notes

Core claim

Under the proposed pattern alternation, the estimated channel at the user is a complex Gaussian random variable in both patterns: the coherent RIS areas contribute a deterministic mean, while the random-phase areas (elements configured for other users) and noise contribute Gaussian spread. Because the real and imaginary parts are independent with equal variance, the squared magnitude of the estimated channel follows a non-central chi-squared distribution with two degrees of freedom, whose non-centrality parameters are the squared magnitudes of the coherent sums and whose variances are the sums of the random-area element powers plus noise. This yields closed-form expressions for the distributions under both hypotheses, and therefore an analytic expression for the detection error probability. The paper verifies these formulas by simulation, showing that the empirical and analytical CDFs match closely, and it documents how the error probability and the power loss vary with the dynamic-part ratio and the element split among users.

Load-bearing premise

The scheme's predicted error rates hold only if the user equipment knows the two expected channel powers needed to set the threshold, and if the RIS element amplitudes are similar enough that the central limit theorem turns the random-phase sums into Gaussian noise.

Editorial extensions

If this is right

With the closed-form CDFs, a system designer can set the detection threshold to minimize error for any given allocation of RIS elements among the coherent, other-user, and dynamic areas, without Monte Carlo simulation.
The scheme trades communication power for sensing accuracy: raising the dynamic-part ratio R lowers the detection error probability but also lowers the average received power for the target UE, so R can be tuned to meet a sensing error target within the link budget.
The detection error probability depends mainly on the size of the dynamic part, not on the split of the remaining elements; increasing the static other-user area mainly raises the relative power difference rather than improving detection.
Because the RIS-assisted path is the only path whose power alternates with the dynamic pattern, the same power measurements both identify the path and separate it from NLOS multipath, a prerequisite for RIS-aided positioning.

Reading between the lines

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

The paper assumes the user knows the two expected power levels needed to set the detection threshold; in practice these would be estimated from limited samples or signaled, and the threshold's robustness to estimation error is an open question the paper leaves implicit.
The central-limit-theorem assumption of similar element amplitudes breaks down in near-field or sparse-aperture settings where pathloss varies strongly across the RIS; a finite-element model or a Gamma approximation would be a testable refinement.
The alternating-pattern idea could be extended to more than two patterns, or to multiple users served in time slots, at the cost of detection latency and power loss; the paper's framework would need a scheduling model to evaluate that trade-off.
Because the scheme requires only power measurements, it could be layered onto existing pilot-based channel estimation with no additional protocol messages, making it a light-weight add-on to RIS-aided localization.

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

Short version: the paper ships a clean statistical derivation for detecting which of two RIS configurations is active, but the motivating scenario—distinguishing the RIS-assisted path from an NLOS path—is never actually in the model. The received signal in Eq. (1) has no NLOS term, and the hypothesis test is between two RIS patterns. That gap is load-bearing, not a footnote.

What is new and good: the idea of using a dynamic RIS sub-surface to amplitude-modulate the estimated channel power is simple and practical. The non-central chi-squared characterization of the two pattern hypotheses (eqs. 19–22) is derived carefully from stated assumptions, the CLT-based Gaussian approximation is standard, and the simulation matches the analysis well (max error-rate difference about 0.01–0.012, max power-difference error about 0.23 dB). No parameters are fitted to the target curves. The trade-off study between the dynamic-part ratio and communication power is exactly the kind of engineering insight a system designer wants. That is real, reproducible work.

Where it softens: (1) The advertised 'distinguish RIS vs NLOS' capability is not demonstrated. If an NLOS component adds to the RIS channel, the observed power is |h_RIS + h_NLOS + n|^2, and the distributions (19)–(22) no longer hold; the noncentrality and variance parameters change. The paper either needs to include an NLOS term in the model (and show how the detector behaves as the NLOS-to-RIS ratio varies) or explicitly state that the UE first resolves paths in delay/angle and that the NLOS case is left to future work. As written, it only demonstrates detecting which pattern is active in the absence of competing paths. (2) The threshold in (7) requires mu1 and mu2, but the paper does not say how the UE learns them. Estimation error on those means is unaddressed; minor-to-moderate, because the UE could plausibly estimate them from the two observed power levels. (3) The similar-amplitude assumption for the CLT is not quantified; for small dynamic parts or strong amplitude variation the Gaussian approximation degrades. Minor. (4) No comparison against alternative path-identification methods; acceptable for a first study.

Net: the paper's central statistical claim holds up on its own terms, but the paper overclaims the application. For a serious referee, it deserves review: the core idea is citable, the analysis is reproducible, and the gap is fixable with a revised model or a reframed claim. I would not cite it for the NLOS-distinction claim as is; I might cite it for the pattern-alternation statistical model once the NLOS issue is addressed.

Recommendation: engage, but require the authors to either model NLOS explicitly or narrow the title and claims.

Referee Report

3 major / 5 minor

Summary. The paper proposes a RIS configuration scheme in which a subset of RIS elements (the dynamic part) is alternately configured for coherent and random combining, so that the power of the estimated channel at a UE is modulated over time. The UE is assumed to detect which of the two RIS patterns is active, based on a threshold test on the estimated channel power. The authors derive the distributions of the estimated channel power under the two patterns as (scaled) non-central chi-squared distributions (Eqs. (19)-(22)), evaluate the detection error probability (Eq. (6)) and relative power difference (Eq. (9)), and investigate the impact of the dynamic-part ratio R and the number of area-2 elements M. Simulation results are reported to match the analytical curves closely, with maximum detection error probability differences around 0.01-0.012.

Significance. If the stated scenario were fully modeled, the paper would offer a simple, low-complexity method for a UE to recognize a RIS-assisted path by exploiting a time-varying power pattern, with an analytical performance evaluation that does not fit parameters to the simulation curves. The non-central chi-squared derivation is internally consistent, and the simulations verify the analysis under the stated assumptions. However, the advertised central claim---identifying RIS-assisted paths in the presence of NLOS paths---is not actually analyzed, because no NLOS path is present in the system model or hypothesis test. The paper's genuine contribution is a detection test between two RIS configurations under an ideal-RIS, no-competitor-path model. The result is plausible and useful within that narrower scope, but the broader claim requires either explicit NLOS modeling or a careful restriction of the claims to pattern detection.

major comments (3)

[Section II.B, Eq. (1)] The system model in Eq. (1) contains only the RIS path, y = h_r^T Omega h_t x + m, with no NLOS scattering component. The hypothesis test in Section II.B is between two RIS configurations (pattern 1 and pattern 2), not between a RIS-assisted path and an NLOS path. The paper's abstract and Section II.B claim that the UE can distinguish the RIS-assisted path from other NLOS paths, but this scenario is never modeled. If an NLOS path is present, the estimated channel would include h_NLOS, so the observed power would be |h_RIS + h_NLOS + n|^2, and the non-central chi-squared distributions in Eqs. (19)-(22) would no longer describe the observed power because their noncentrality parameters and variances omit the NLOS term. The central claim therefore is not demonstrated; either the NLOS term must be added to the model and the analysis re-derived, or the paper must be reframed as a pattern-detection scheme for a known RIS path in the absence of competing paths.
[Section II.B, Eq. (7)] The threshold gamma in Eq. (7) is defined as the minimizer of the detection error probability and depends on mu1 and mu2, the expected powers under the two hypotheses. Those expectations depend on the channel gains of all RIS paths and on the noise variance, as seen in Eqs. (19)-(22). The paper does not state how a UE obtains mu1 and mu2 or gamma in practice. Without a concrete estimation or training procedure, Eq. (6) describes an oracle threshold rather than an achievable detection error probability. The authors should specify how the threshold is set (e.g., estimated from pilot measurements, derived from known geometry, or chosen robustly) and evaluate the impact of threshold mismatch.
[Section III.A, Eq. (12)] The Gaussian approximation in Eq. (12) relies on the assumption that all channel coefficients have similar amplitudes and that M is sufficiently large. In a realistic RIS scenario, channel amplitudes vary with distance and angle, and the phases of the random-combining elements are not necessarily i.i.d. uniform over 2pi if the physical configuration is constrained. The paper does not quantify the validity range of this approximation beyond the single simulation setup. Since this assumption is load-bearing for the non-central chi-squared result, the authors should provide a robustness study or an explicit discussion of the conditions under which the approximation degrades, particularly for small M or K and for non-uniform amplitudes.

minor comments (5)

[Section IV.C] The phrase "respsctivly" in the description of Fig. 4 is a typo; it should be "respectively."
[Appendix A] In Eq. (28), the summation index is written as N although the dynamic part hd is defined with K elements in Eq. (23). This is confusing and should be corrected to K.
[Section IV.B, Fig. 3] The caption of Fig. 3 is unclear: "alpha = 1/sigma_1^2, h = h_1" does not explain the plotted quantity well. It should state explicitly that the figure plots the empirical and analytical CDFs of the scaled power 1/sigma_i^2 |h_i|^2 for i = 1,2.
[References] Reference [6] is a duplicate of reference [1]; the duplicate should be removed or replaced with a distinct citation.
[Section IV.C] The text says "the relative power slightly increases" and later "the relative power difference increases," but Fig. 4 shows the relative power difference as a function of R. The wording is imprecise and should be rephrased to avoid confusion.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the analysis is derived from stated statistical assumptions and verified by simulation, with no fitted parameter or load-bearing self-citation.

full rationale

The paper's derivation chain is self-contained. The received signal model (1) and the two hypotheses in Section II.B define the detection problem directly in terms of the two RIS patterns. The statistical model in Section III derives the estimated channel power distributions (19)-(20) from the central limit theorem and from explicit variance/correlation computations in Appendices A-C, with stated assumptions (uniform random phases, similar amplitudes, independent noise). The detection error probability (6) is then evaluated analytically from these distributions, with the threshold gamma chosen as the minimizer in (7). No parameter is fitted to the simulation curves; the simulations implement the same model and confirm the algebra. The only self-citations are [7] and [8], which appear in the introductory paragraph about RIS shaping the propagation environment and do not enter the derivations. The paper's advertised NLOS-identification scenario is not fully modeled: Eq. (1) contains no NLOS component, and the H1/H2 test distinguishes two RIS patterns rather than RIS versus NLOS. However, this is a scope/modeling limitation, not a circular reduction: the paper does not define its detection result in terms of NLOS power, nor does it fit any parameter to make the error probability match. The practical availability of the expected powers mu1 and mu2 for threshold setting is an implementation question, not a circularity concern. Therefore no circular step is exhibited.

Assumptions & free parameters 0 free parameters · 4 assumptions · 0 invented entities

The analysis rests on standard CLT-based statistical modeling; no free parameters are fitted to data, and no new physical entities are introduced. The simulation parameters (N, M, K, R) are scenario inputs, not fitted values.

assumptions (4)

domain assumption Phase shifts of area-2 and area-3 (in pattern 2) are independent and uniformly distributed on (-pi, pi)
Used in Section III.A, eqs. (12) and (14), to model the sums as zero-mean complex Gaussian via the CLT.
domain assumption All channel coefficients hr,q and ht,q have similar amplitude
Stated in Section III.A to justify the Gaussian approximation of the random-phase sums; also used implicitly in the variance expressions.
domain assumption Free-space pathloss channel model with ideal RIS (no amplitude attenuation, full 2pi phase control)
Used in eqs. (5) and (10); idealizes the RIS phase response and channel model.
standard math Central limit theorem applies to the sum of independent random-phase contributions with a sufficient number of elements
Invoked in Section III.A to replace the random-phase sums by complex Gaussian variables.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Identification of RIS-Assisted Paths for Wireless Integrated Sensing and Communication." pith.science (2026). https://pith.science/paper/WAXRCEWG

@misc{pith2026250604123,
  author       = {Pith},
  title        = {Pith review of: Identification of RIS-Assisted Paths for Wireless Integrated Sensing and Communication},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/WAXRCEWG}},
  note         = {Machine review of arXiv:2506.04123}
}

read the original abstract

Distinguishing between reconfigurable intelligent surface (RIS) assisted paths and non-line-of-sight (NLOS) paths is a fundamental problem for RIS-assisted integrated sensing and communication. In this work, we propose a pattern alternation scheme for the RIS response that uses part of the RIS as a dynamic part to modulate the estimated channel power, which can considerably help the user equipments (UEs) to identify the RIS-assisted paths. Under such a dynamic setup, we formulate the detection framework for a single UE, where we develop a statistical model of the estimated channel power, allowing us to analytically evaluate the performance of the system. We investigate our method under two critical factors: the number of RIS elements allocated for the dynamic part and the allocation of RIS elements among different users. Simulation results verify the accuracy of our analysis.

Figures

Figures reproduced from arXiv: 2506.04123 by the authors.

**Figure 2.** Scenario of our system model. hRIS = [PITH_FULL_IMAGE:figures/full_fig_p002_2.png] view at source ↗

**Figure 3.** Empirical and analytical CDFs. In pattern 1, [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗

Figures from the paper (1 more)

**Figure 5.** Figure 5: Detection error probability and relative power loss w.r.t. [PITH_FULL_IMAGE:figures/full_fig_p005_5.png]

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

13 extracted references · 10 canonical work pages

[1]

De Lima et al., ”Convergent Communication, Sensing and Localiza- tion in 6G Systems: An Overview of Technologies, Opportunities and Challenges,” in IEEE Access, vol

C. De Lima et al., ”Convergent Communication, Sensing and Localiza- tion in 6G Systems: An Overview of Technologies, Opportunities and Challenges,” in IEEE Access, vol. 9, pp. 26902-26925, 2021

work page 2021
[2]

Y . Cui, F. Liu, X. Jing and J. Mu, ”Integrating Sensing and Commu- nications for Ubiquitous IoT: Applications, Trends, and Challenges,” in IEEE Network, vol. 35, no. 5, pp. 158-167, September/October 2021

2021
[3]

Bresson, Z

G. Bresson, Z. Alsayed, L. Yu and S. Glaser, ”Simultaneous Localization and Mapping: A Survey of Current Trends in Autonomous Driving,” in IEEE Transactions on Intelligent Vehicles, vol. 2, no. 3, pp. 194-220, Sept. 2017, doi: 10.1109/TIV .2017.2749181

arXiv 2017
[4]

Bartoletti, A

S. Bartoletti, A. Conti and M. Z. Win, ”Device-free counting via wideband signals”, IEEE J. Sel. Areas Commun., vol. 35, no. 5, pp. 1163-1174, May 2017

work page 2017
[5]

M. Di Renzo et al., ”Smart Radio Environments Empowered by Recon- figurable Intelligent Surfaces: How It Works, State of Research, and The Road Ahead,” in IEEE Journal on Selected Areas in Communications, vol. 38, no. 11, pp. 2450-2525, Nov. 2020

work page 2020
[6]

De Lima et al., ”Convergent Communication, Sensing and Local- ization in 6G Systems: An Overview of Technologies, Opportunities and Challenges,” in IEEE Access, vol

C. De Lima et al., ”Convergent Communication, Sensing and Local- ization in 6G Systems: An Overview of Technologies, Opportunities and Challenges,” in IEEE Access, vol. 9, pp. 26902-26925, 2021, doi: 10.1109/ACCESS.2021.3053486

work page arXiv 2021
[7]

Tahir, S

B. Tahir, S. Schwarz and M. Rupp, ”Analysis of Uplink IRS-Assisted NOMA Under Nakagami-m Fading via Moments Matching,” in IEEE Wireless Communications Letters, vol. 10, no. 3, pp. 624-628, March 2021, doi: 10.1109/LWC.2020.3043810

work page arXiv 2021
[8]

L. Hao, S. Schwarz and M. Rupp, ”Analysis and Optimization of Reconfigurable Intelligent Surfaces Assisted MIMO Systems,” 2021 Joint European Conference on Networks and Communications & 6G Summit (EuCNC/6G Summit), 2021, pp. 13-18, doi: 10.1109/Eu- CNC/6GSummit51104.2021.9482594

arXiv 2021

Show all 13 references

[9]

Reconfigurable intelligent sur- faces for localization: Position and orientation error bounds[J]

Elzanaty A, Guerra A, Guidi F, et al. Reconfigurable intelligent sur- faces for localization: Position and orientation error bounds[J]. IEEE Transactions on Signal Processing, 2021, 69: 5386-5402

2021
[10]

RIS-Enabled Localization Con- tinuity Under Near-Field Conditions[C]//2021 IEEE 22nd International Workshop on Signal Processing Advances in Wireless Communications (SPAWC)

Rahal M, Denis B, Keykhosravi K, et al. RIS-Enabled Localization Con- tinuity Under Near-Field Conditions[C]//2021 IEEE 22nd International Workshop on Signal Processing Advances in Wireless Communications (SPAWC). IEEE, 2021: 436-440

2021
[11]

Albanese, P

A. Albanese, P. Mursia, V . Sciancalepore and X. Costa-P ´erez, ”PAPIR: Practical RIS-aided Localization via Statistical User Information,” 2021 IEEE 22nd International Workshop on Signal Processing Advances in Wireless Communications (SPAWC), 2021, pp. 531-535

2021
[12]

J. He, H. Wymeersch, L. Kong, O. Silv ´en and M. Juntti, ”Large Intelli- gent Surface for Positioning in Millimeter Wave MIMO Systems,”IEEE 91st Vehicular Technology Conference (VTC2020-Spring), 2020

2020
[13]

S. Hu, F. Rusek and O. Edfors, ”Beyond Massive MIMO: The Potential of Positioning With Large Intelligent Surfaces,” in IEEE Transactions on Signal Processing, vol. 66, no. 7, pp. 1761-1774, 1 April1, 2018, doi: 10.1109/TSP.2018.2795547. APPENDIX A. Statistics of the dynamic pa...

2018

Pith tools

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

[1] [1]

De Lima et al., ”Convergent Communication, Sensing and Localiza- tion in 6G Systems: An Overview of Technologies, Opportunities and Challenges,” in IEEE Access, vol

C. De Lima et al., ”Convergent Communication, Sensing and Localiza- tion in 6G Systems: An Overview of Technologies, Opportunities and Challenges,” in IEEE Access, vol. 9, pp. 26902-26925, 2021

work page 2021

[2] [2]

Y . Cui, F. Liu, X. Jing and J. Mu, ”Integrating Sensing and Commu- nications for Ubiquitous IoT: Applications, Trends, and Challenges,” in IEEE Network, vol. 35, no. 5, pp. 158-167, September/October 2021

2021

[3] [3]

Bresson, Z

G. Bresson, Z. Alsayed, L. Yu and S. Glaser, ”Simultaneous Localization and Mapping: A Survey of Current Trends in Autonomous Driving,” in IEEE Transactions on Intelligent Vehicles, vol. 2, no. 3, pp. 194-220, Sept. 2017, doi: 10.1109/TIV .2017.2749181

arXiv 2017

[4] [4]

Bartoletti, A

S. Bartoletti, A. Conti and M. Z. Win, ”Device-free counting via wideband signals”, IEEE J. Sel. Areas Commun., vol. 35, no. 5, pp. 1163-1174, May 2017

work page 2017

[5] [5]

M. Di Renzo et al., ”Smart Radio Environments Empowered by Recon- figurable Intelligent Surfaces: How It Works, State of Research, and The Road Ahead,” in IEEE Journal on Selected Areas in Communications, vol. 38, no. 11, pp. 2450-2525, Nov. 2020

work page 2020

[6] [6]

De Lima et al., ”Convergent Communication, Sensing and Local- ization in 6G Systems: An Overview of Technologies, Opportunities and Challenges,” in IEEE Access, vol

C. De Lima et al., ”Convergent Communication, Sensing and Local- ization in 6G Systems: An Overview of Technologies, Opportunities and Challenges,” in IEEE Access, vol. 9, pp. 26902-26925, 2021, doi: 10.1109/ACCESS.2021.3053486

work page arXiv 2021

[7] [7]

Tahir, S

B. Tahir, S. Schwarz and M. Rupp, ”Analysis of Uplink IRS-Assisted NOMA Under Nakagami-m Fading via Moments Matching,” in IEEE Wireless Communications Letters, vol. 10, no. 3, pp. 624-628, March 2021, doi: 10.1109/LWC.2020.3043810

work page arXiv 2021

[8] [8]

L. Hao, S. Schwarz and M. Rupp, ”Analysis and Optimization of Reconfigurable Intelligent Surfaces Assisted MIMO Systems,” 2021 Joint European Conference on Networks and Communications & 6G Summit (EuCNC/6G Summit), 2021, pp. 13-18, doi: 10.1109/Eu- CNC/6GSummit51104.2021.9482594

arXiv 2021