Pith. sign in

REVIEW 4 major objections 5 minor 14 references

Performance Analysis of BEM-based Channel Estimation for OTFS with Hardware Impairments

T0 review · 4 major / 5 minor · reviewed 2026-08-08 · deepseek-v4-flash

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

Pith's one-line read Hardware impairments sharply degrade OTFS channel estimation and BER, with closed-form MSE and lower-bound expressions.

desk verdict MSE part is solid and useful; the BER lower bound has an internal channel-model inconsistency that should be fixed before it is cited. read the letter →

arxiv 2502.04003 v1 pith:NVGAYDBI submitted 2025-02-06 eess.SP

Haowei Wu , Huanyu Chen , Qihao Peng , Qu Luo , Jinglan Ou This is my paper · ORCID
classification eess.SP MSC 94A1294A40
keywords OTFSchannelestimationbasisexpansionmodelhardwareimpairmentsMMSEdetectorbiterrorratemeansquaredelay-Dopplerdomain
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

This letter tries to establish that OTFS systems are sensitive to residual hardware impairments and that this sensitivity is exactly quantifiable. Using a basis expansion model for the time-varying channel, the authors derive a closed-form mean square error expression for MMSE channel estimation that includes transmitter and receiver quality factors, distortion noise, and basis expansion modeling error. They then derive a lower bound on the average bit error rate of an MMSE detector from per-symbol SINR expressions. The formulas show that even small departures from ideal hardware noticeably degrade error performance, especially at high SNR where thermal noise no longer dominates. A sympathetic reader should care because the results set a concrete hardware-quality target for OTFS to deliver its Doppler-resilience advantage.

What carries the argument

The central machinery is the generalized complex exponential basis expansion model (GCE-BEM), which represents each time-varying channel tap as a small sum of exponential basis functions plus a modeling error term, reducing the channel to a coefficient vector of size $(Q_L+1)(L+1)$. Hardware impairments enter through two scalar quality factors $\xi_i$ and $\xi_o$ that scale the useful signal and add Gaussian distortion noise at the transmitter and receiver. An MMSE estimator for the BEM coefficients and an MMSE detector in the time domain carry the analysis; their error covariance matrices and per-symbol SINR expressions produce the MSE formula (13) and the BER lower bound (33).

What would settle it

Measure a real OTFS transceiver's MSE and BER against equations (13) and (33) across SNR values while independently measuring the hardware quality factors; if the high-SNR error floor departs from the predicted floor, or moves with modulation format in a way (33) cannot capture, the Gaussian distortion model is falsified.

Watch

Extended reading notes

Core claim

The paper claims that for an OTFS link whose transmitter and receiver suffer residual hardware impairments, modeled as independent zero-mean Gaussian distortion noises with covariances $(1-\xi_i)E[|s|^2]I$ and $(1-\xi_o)E[|H_t s_i|^2]I$, the MSE of the BEM-based MMSE channel estimator is given by (13) and the average BER of the MMSE detector is lower-bounded by (33). The MSE expression splits the estimation error into BEM coefficient error plus modeling error, while the BER bound follows by applying Jensen's inequality to the per-symbol SINR. The paper validates both formulas by simulation and concludes that even minor hardware impairments cause significant performance degradation, most visibly in the high-SNR regime where AWGN is no longer the dominant error source.

Load-bearing premise

The central assumption is that residual hardware impairments are independent zero-mean Gaussian distortion noises whose powers scale with the signal via the quality factors; if real impairments are nonlinear, non-Gaussian, or correlated with the signal, the derived MSE and BER formulas break down.

Editorial extensions

If this is right

  • Equation (13) lets system designers predict channel estimation accuracy from transceiver quality factors alone, without running Monte Carlo simulations.
  • The BER lower bound (33) exposes an SNR ceiling set by hardware quality: beyond it, increasing transmit power stops improving the error rate.
  • The analysis identifies the high-SNR region as the regime where hardware impairments dominate, so OTFS performance there is limited by hardware quality rather than by thermal noise.
  • The BEM-based estimator keeps pilot overhead low (7.5% in the simulations) while making the hardware-impaired MSE analytically tractable.

Reading between the lines

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

  • If the Gaussian distortion model were replaced by measured impairment statistics, such as IQ imbalance or amplifier nonlinearity, the same derivation structure could produce adjusted MSE and BER expressions, but the paper does not test such distributions.
  • Equation (33) suggests a practical hardware calibration target: transceivers need quality factors above some threshold before OTFS Doppler resilience is realized, and extracting that threshold from the bound is a natural follow-up.
  • The simulations' comparison with OFDM implies the framework could be extended to quantify hardware-impairment penalties across modulation waveforms, though the paper only gestures at that comparison.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

4 major / 5 minor

Summary. This letter analyzes BEM-based channel estimation and MMSE detection for OTFS systems under transceiver hardware impairments, modeled through quality factors xi_i and xi_o with Gaussian distortion noises. The paper derives the MMSE channel estimator and a closed-form MSE expression (Eq. 13), then proposes a time-domain MMSE detector and derives a lower bound on the average BER (Eq. 33). Numerical simulations are presented to validate the MSE and BER analyses.

Significance. If the derivations were sound, the paper would provide a useful extension of BEM-OTFS channel estimation to practical hardware-impaired transceivers; the MSE analysis is structurally standard and the Fig. 3 match between simulation and theory supports that portion. The BER analysis, however, is not supported by the derivation as written due to inconsistent use of the estimated versus true channel and algebraic errors in Appendix B. The qualitative conclusion that hardware impairments degrade BER may still be correct, but the theoretical backing needs substantial repair.

major comments (4)
  1. [Appendix B, Eqs. (28)-(29)] The first displayed expression for \hat x_d[i] in Eq. (28) does not follow from Eq. (15). Since (15) gives \hat r_o = sqrt(xi_o xi_i) \hat H_t s_d + n, with n containing sqrt(xi_o xi_i) \tilde H_t s, the DD-domain estimate should be \hat x_d = sqrt(xi_o xi_i) F G_t \hat H_t F^H x_d + F G_t n. Replacing \hat H_t by H_t in (28) and in T = sqrt(xi_o xi_i) F G_t H_t F^H double-counts the channel estimation error, because n already includes the \tilde H_t s term. The missing cross terms sqrt(xi_o xi_i) G_t( \hat H_t R_sd \tilde H_t^H + \tilde H_t R_sd \hat H_t^H ) G_t^H do not vanish in general. Consequently, the SINR expression (31) and the lower bound (33) are not derived from the stated model.
  2. [Appendix B, Eq. (29)] The algebra in Eq. (29) is incorrect: the noise term from the first line is f_i G_t R_n G_t^H f_i^H, so the second line should be f_i G_t( xi_o xi_i H_t R_sd H_t^H + R_n ) G_t^H f_i^H, not xi_o xi_i f_i G_t( H_t R_sd H_t^H + R_n ) G_t^H f_i^H. As a result, the equality to xi_o xi_i sigma_d^2 T[i,i] does not hold dimensionally or algebraically; with T defined as in the paper, the factor of xi_o xi_i is inconsistent.
  3. [Appendix B, Eqs. (17) and (33)] The denominator of the averaged T[i,i] term is written as N, but N is the Doppler dimension of the OTFS grid; the average over the Nnum data symbols should be divided by Nnum. As written, the lower bound is not well-defined, and the numerical value of the bound would depend on this normalization. This should be corrected and the derivation should state which quantity is actually averaged.
  4. [Appendix B, Eq. (33) and R_n definition] The Jensen step requires convexity of eta(x) = erfc( sqrt(b_M x/(1-x)) ) on (0,1), which is asserted by citation [14] but not proven or stated as a lemma in this paper; since this is load-bearing for the lower bound, the authors should provide a proof or a precise statement of the convexity condition for the b_M used. In addition, the paper never gives an explicit expression for R_n = E[n n^H] used in G_t and T; without it, the theoretical BER curve cannot be reproduced from the equations, and the paper should state whether R_n is computed analytically from (14) or estimated from simulation.
minor comments (5)
  1. [Appendix A] The first sentence contains a duplicated word: "with with zero mean" should be "with zero mean".
  2. [Section IV] "exhibits a increase in volatility" should read "exhibits an increase in volatility".
  3. [Figure 2 caption] "Pilot Patten" is a typo for "Pilot Pattern".
  4. [Section III-B, after Eq. (10)] The sentence "Rhh denotes the covariance matrix of the BEM coefficient vectors and the CIR vectors" is ambiguous; it should clarify that Rhh is the covariance of the CIR vector h and Rc is the covariance of the BEM coefficient vector c.
  5. [Theorem 2, Eq. (32)] The summation index starts at i=0 in Eq. (32) but at i=1 in Eq. (33); the indexing should be made consistent, presumably summing over i=1,...,Nnum.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: the MSE and BER derivations are self-contained from the stated impairment and BEM models, and the simulations are consistency checks against the same model rather than fitted predictions.

full rationale

No circular step was found. The channel-estimation MSE in (13) follows from the MMSE estimator in (9) together with the noise-plus-interference covariance in (14), which is assembled from the stated Gaussian hardware-impairment model in (1)-(2), the GCE-BEM representation in (4)-(5), and the data/pilot correlation statistics. No quantity in the MSE or BER expression is fitted to the simulation curves; theoretical curves are derived from stated statistical assumptions and then compared with simulations generated under the same model, which is a standard consistency check rather than a circular prediction. The BER lower bound in (33) is obtained by applying Jensen's inequality to the per-symbol SINR in (31), following the external derivation approach of [12] and using the convexity of the erfc-based function from [14]; neither [12] nor [14] is authored by the present authors, and neither citation is load-bearing in a self-referential way. The paper contains no self-citation chain and does not rename a known result as a new prediction. One internal algebraic concern exists in Appendix B: Eq. (28) writes the desired-signal term with H_t while n in Eq. (15) already contains the term sqrt(xi_o xi_i) tilde-H_t s, so the power identity in (29) may double-count the channel-estimation error. However, this is a derivation-error or correctness issue, not a circular reduction of the conclusion to its own inputs, and it does not affect the circularity score under the given criteria.

Assumptions & free parameters 1 free parameters · 5 assumptions · 0 invented entities

The analysis rests on standard statistical models from prior work: the Gaussian residual hardware impairment model, the Jakes' channel covariance, and the GCE-BEM decomposition with known modeling-error covariance. The only hand-chosen design parameter is the BEM frequency resolution factor R=2. No new entities are introduced and no constants are fitted to the simulation data.

free parameters (1)
  • BEM frequency resolution factor R = 2
    Chosen in Section II-B as a complexity versus modeling-error trade-off following [5]; the derivations hold for general R, but simulations only validate R=2. It is a hand-set design parameter, not fitted to data.
assumptions (5)
  • domain assumption Residual hardware impairments are zero-mean complex Gaussian distortion noises independent of the signal, with variances (1-xi_i)E|s|^2 and (1-xi_o)E|H_t s_i|^2.
    Equations (1)-(2), citing [10] and [11]; this model underpins all subsequent MSE and BER expressions. Real impairments can be non-Gaussian and signal-dependent.
  • domain assumption The receiver knows the channel second-order statistics R_hh and the impairment and noise variances needed to form the MMSE estimator in (9) and detector in (16).
    Used throughout Section III; the paper does not discuss how these statistics are acquired or the cost of imperfect knowledge.
  • domain assumption The GCE-BEM modeling error e_mod is orthogonal to the BEM coefficient subspace, and its covariance is E[e_mod e_mod^H] = G R_h,l' G as in [5].
    Used in (11)-(13) and Appendix A; imported from [5] without proof in this paper.
  • standard math eta(x) = erfc(sqrt(b_M x/(1-x))) is convex on x in (0,1), cited from [14] and used for Jensen's inequality.
    Load-bearing for the BER lower bound (33); not proved in this paper, only cited.
  • domain assumption Data symbols are i.i.d. with power sigma_d^2 and independent of the BEM coefficients and channel, so interference can be summarized by the covariance Rz in (14).
    Needed for the linear MMSE channel estimator and the SINR computation.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Performance Analysis of BEM-based Channel Estimation for OTFS with Hardware Impairments." pith.science (2026). https://pith.science/paper/NVGAYDBI

@misc{pith2026250204003,
  author       = {Pith},
  title        = {Pith review of: Performance Analysis of BEM-based Channel Estimation for OTFS with Hardware Impairments},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/NVGAYDBI}},
  note         = {Machine review of arXiv:2502.04003}
}
read the original abstract

This letter studies the low-complexity channel estimation for orthogonal time frequency space (OTFS) in the presence of hardware impairments. Firstly, to tackle the computational complexity of channel estimation, the basis expansion model (BEM) is utilized. Then, the mean square error (MSE) of the estimated channel is theoretically derived, revealing the effects of hardware impairments on channel estimation. Based on the estimated channel, the minimum mean square error (MMSE) detector is adopted to analyze the impacts of imperfect hardware on the bit error rate (BER). Finally, the numerical results validate the correctness of our theoretical analysis of the MSE for channel estimation and lower bound of the BER, and also demonstrate that even minor hardware impairments can significantly degrade the performance of the OTFS system.

Figures

Figures reproduced from arXiv: 2502.04003 by the authors.

Figure 1
Figure 1. Hardware Impairment OTFS System Model. A. Hardware Impairment OTFS Signal Model The block diagram of the OTFS system in the presence of hardware impairments is shown in [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. OTFS System Pilot Patten. A. Pilot Design As illustrated in [PITH_FULL_IMAGE:figures/full_fig_p002_2.png] view at source ↗
Figure 3
Figure 3. MSE performance with different quality factors. (a) [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗
Figures from the paper (1 more)
Figure 4
Figure 4. Figure 4: BER performance with different quality factors. (a) [PITH_FULL_IMAGE:figures/full_fig_p005_4.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

14 extracted references · 14 canonical work pages

  1. [14]

    Z. Li, C. Zhang, G. Song, X. Fang, X. Sha and D. Slock, ”Chirp Parameter Selection for Affine Frequency Division Multiplexing with MMSE Equalization,” IEEE Trans. Commun

    work page

  2. [1]

    X. He, W. Yuan and P. Fan, ”On the pilot-aided channel estimation for windowed OTFS with data interference in rapidly time-varying channels,” IEEE Trans. Wireless Commun. , vol. 23, no. 11, pp. 16359-16374, Nov. 2024

    work page 2024

  3. [2]

    J. Wu, W. Yuan, Z. Wei, K. Zhang, F. Liu and D. W. K. Ng, ”Low- complexity minimum BER precoder design for ISAC systems: A delay- Doppler perspective,” IEEE Trans. Wireless Commun

    work page

  4. [3]

    Hadani et al., ”Orthogonal time frequency space modulation,” in Proc

    R. Hadani et al., ”Orthogonal time frequency space modulation,” in Proc. IEEE Wireless Commun. Netw. Conf. (WCNC) , Mar. 2017, pp. 1-6

    work page 2017

  5. [4]

    Raviteja, K

    P. Raviteja, K. T. Phan and Y . Hong, ”Embedded pilot-aided channel estimation for OTFS in delay–Doppler channels,” IEEE Trans. Veh. Technol., vol. 68, no. 5, pp. 4906-4917, May 2019

    work page 2019

  6. [5]

    Y . Liu, Y . L. Guan and D. G. G., ”Near-optimal BEM OTFS receiver with low pilot overhead for high-mobility communications,” IEEE Trans. Commun., vol. 70, no. 5, pp. 3392-3406, May 2022

    work page 2022

  7. [6]

    Y . Liu, Y . L. Guan and D. G. G., ”Turbo BEM OTFS receiver with optimized superimposed pilot power,” IEEE Trans. Commun. , vol. 72, no. 1, pp. 601-617, Jan. 2024

    work page 2024

  8. [7]

    Tusha, S

    A. Tusha, S. Do ˘gan-Tusha, F. Yilmaz, S. Althunibat, K. Qaraqe and H. Arslan, ”Performance analysis of OTFS under in-phase and quadrature imbalance at transmitter,” IEEE Trans. Veh. Technol., vol. 70, no. 11, pp. 11761-11771, Nov. 2021. JOURNAL OF LATEX CLASS FILES, VOL. 14, NO. 8, AUGUST 2021 7

    work page 2021

Show all 14 references
  1. [8]

    S. G. Neelam and P. R. Sahu, ”Analysis, estimation and compensation of hardware impairments for CP-OTFS systems,” IEEE Wireless Commun. Letters, vol. 11, no. 5, pp. 952-956, May 2022

    2022

  2. [9]

    Singh, S

    A. Singh, S. Sharma, M. Sharma, K. Deka and D. B. da Costa, ”Autoencoder-based end-to-end OTFS system design with hardware im- pairments,” IEEE Wireless Commun. Letters , vol. 13, no. 8, pp. 2285- 2289, Aug. 2024

    2024

  3. [10]

    X. Li, J. Li, Y . Liu, Z. Ding and A. Nallanathan, ”Residual transceiver hardware impairments on cooperative NOMA networks,” IEEE Trans. Wireless Commun., vol. 19, no. 1, pp. 680-695, Jan. 2020

    2020

  4. [11]

    Tubail, B

    D. Tubail, B. Ceniklioglu, A. E. Canbilen, I. Develi and S. S. Ikki, ”The effect of hardware impairments on the error bounds of Localization and maximum likelihood estimation of mm-Wave MISO-OFDM systems,” IEEE Trans. Veh. Technol., vol. 72, no. 3, pp. 4063-4067, March 2023

    2023

  5. [12]

    Signal Processing, vol

    Yuan-Pei Lin and See-May Phoong, ”BER minimized OFDM systems with channel independent precoders,” IEEE Trans. Signal Processing, vol. 51, no. 9, pp. 2369-2380, Sept. 2003

    2003

  6. [13]

    Goldsmith, Wireless communications

    A. Goldsmith, Wireless communications . Cambridge university press, 2005

    2005

Pith tools

Reviewed August 8, 2026 · model on record in the stance chip above.