Pith. sign in

REVIEW 3 major objections 6 minor 12 references

Channel Spreading Function-Inspired Channel Transfer Function Estimation for OFDM Systems with High-Mobility

T0 review · 3 major / 6 minor · reviewed 2026-08-11 · deepseek-v4-flash

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

Pith's one-line read Under compact delay-Doppler conditions, one period of a pilot-derived periodic CSF recovers the OFDM channel exactly, and the data-symbol CTF follows by inverse transform.

desk verdict Solid on-grid CSF recovery and a useful pilot-spacing rule; off-grid 'exact' claim needs softening to approximate. read the letter →

arxiv 2412.07074 v1 pith:2S54SQVP submitted 2024-12-10 eess.SP

Yiyan Ma , Bo Ai , Guoyu Ma , Akram Shafie , Qingqing Cheng , Mi Yang , Jingli Li , Xuebo Pang
show 2 more authors
Jinhong Yuan Zhangdui Zhong
This is my paper · ORCID
classification eess.SP
keywords OFDMchanneltransferfunctionestimationspreadinghighmobilitydelay-DopplerdomainpilotdesigndoublyselectivefractionalDoppler
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 paper tries to show that an OFDM receiver can obtain the full channel transfer function (CTF) under high mobility from pilot symbols alone, without interpolating across fading. The route is to estimate the channel spreading function (CSF) in the delay-Doppler domain, which stays quasi-static over intervals much longer than the channel coherence time, and then transform it back to the time-frequency domain for every data symbol. The paper proves that, when all path delays lie on the delay grid and the delay and Doppler supports fit inside ranges set by pilot spacing, one period of the periodic CSF computed from pilots exactly equals the original CSF, so no resolution is lost. In simulations the resulting CTF estimator matches the MMSE estimator's accuracy without needing channel statistics and runs at roughly DFT-level complexity. If correct, this gives high-mobility OFDM a practical channel estimator that keeps the current waveform and only changes where pilots sit.

What carries the argument

The central object is the periodic channel spreading function $h_{\mathrm{DD}}^{\mathrm{Periodic}}[k,l]=\mathrm{DFT}_N\{\mathrm{IDFT}_M\{h_{\mathrm{TF}}^{\mathrm{Discrete}}[m,n]\}\}$, computed from pilot-only CTF samples spaced by $d_t$ in time and $d_f$ in frequency. It is two-dimensionally periodic with periods $N/d_t$ and $M/d_f$, so one period reproduces the original CSF exactly when the channel's delay and Doppler supports fit inside those periods; the proof compares the geometric-series kernels $R_{\mathrm{Doppler}}^{\mathrm{Periodic}}(k_i,k)$ and $R_{\mathrm{delay}}^{\mathrm{Periodic}}(l_i,l)$ with their full-grid counterparts and fixes the scaling factors $\tilde{M}=M/d_f^2$, $\tilde{N}=N/d_t^2$. In the off-grid Doppler case, Algorithm 1 locates each path by the peak of this kernel and estimates the fractional part of the Doppler from the magnitude ratio of the two neighboring samples, using the large-$N$ approximation $\sin(\pi d_t x/N)\approx \pi d_t x/N$.

What would settle it

Simulate a two-path channel whose delays fall into the same resolvable delay bin but whose Dopplers differ, with on-grid delays and high SNR; if Algorithm 1 cancels the two paths together and the CTF MSE departs from the MMSE curve, the exact-recovery claim fails in exactly the regime the theorems exclude. A second test simulates continuously random delays and checks whether the MSE gap to MMSE appears precisely when a delay moves off the grid.

Watch

Extended reading notes

Core claim

The central claim is that the delay-Doppler channel spreading function of an OFDM link can be recovered exactly from one period of the periodic CSF obtained by applying a length-$N$ DFT along time and a length-$M$ IDFT along frequency to the CTF sampled only at pilot resource elements. Theorem 1 states that if the normalized Doppler lies in $[-1/(2d_t T),\,1/(2d_t T)-1/(NT)]$ and the normalized delay lies in $[0,\,1/(d_f \Delta f)-1]$, with delays on-grid, then $h_{\mathrm{DD}}^{\mathrm{Periodic}}[k,l]$ exactly represents the original CSF on that period, and the CTF for all data symbols follows from the inverse transform in (16). For off-grid Doppler, Algorithm 1 estimates each path's fractional Doppler from the ratio of the two largest adjacent peaks of the sinc-like kernel, reconstructs the CSF via (21), and Theorem 2 asserts exact recovery under the same compactness bounds. The numerical results show MSE close to the MMSE estimator and BER close to the ideal-channel benchmark in both on-grid and off-grid Doppler cases.

Load-bearing premise

The exact-recovery theorems assume every path's delay lands exactly on the integer delay grid, that at most one path occupies each resolvable delay bin, and that the number of paths is known from an energy threshold.

Editorial extensions

If this is right

  • In high-mobility OFDM links satisfying the compactness bounds, pilot spacing $d_t$, $d_f$ carries enough information to recover every data-symbol CTF exactly, so time-frequency interpolation across fast fading is unnecessary.
  • The estimator's complexity is $O(N\log N + M\log M + P)$, versus the MMSE estimator's $O((N/d_t)^3(M/d_f)^3)$, and it needs no covariance or channel-statistics information.
  • With a single-tap equalizer, the proposed method's BER approaches the ideal-channel benchmark and tracks the MMSE estimator in both on-grid and off-grid Doppler simulations.
  • The MSE is slightly higher in the off-grid Doppler case because of the fractional-Doppler approximation, but the BER penalty is negligible.
  • The pilot arrangement is a standard lattice with $d_t$ and $d_f$ chosen from the CSF compactness conditions, so the approach is compatible with existing OFDM frames.

Reading between the lines

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

  • A practical corollary the paper leaves implicit is a closed-form pilot-spacing rule: choose $d_t$, $d_f$ so that $1/(2d_t T)\geq \nu_{\max}$ and $1/(d_f \Delta f)\geq \tau_{\max}$, which could be adapted per mobility class in a standards implementation.
  • Because the estimated CSF contains path delays and Dopplers, the same pilot snapshot could feed integrated sensing and communication, turning channel estimation into radar parameter estimation without extra pilots.
  • The fractional-Doppler step could be iterated: after each path is subtracted in Algorithm 1, a second peak-ratio pass or a Newton refinement of $\hat{k}_i$ would likely remove the small off-grid MSE gap to MMSE; this is a testable extension, not claimed in the paper.
  • The one-path-per-delay-bin assumption could be relaxed by oversampling the delay axis or applying a subspace method inside each occupied bin; the paper's exact statements do not cover that case.
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

3 major / 6 minor

Summary. The manuscript proposes a CSF-inspired CTF estimation method for OFDM systems in high-mobility scenarios. Pilots placed on a TF lattice are used to form a discrete CTF, whose 2D DFT/IDFT yields a periodic CSF. Theorem 1 states that under on-grid delay and Doppler conditions with compact support (Doppler within [-1/(2d_t T), 1/(2d_t T) - 1/(NT)] and delay within [0, 1/(d_f Δf) - 1]), one period of the pilot-derived periodic CSF exactly equals the original CSF. Theorem 2 claims exact recovery also for off-grid Doppler using Algorithm 1, which estimates fractional Doppler from the ratio of adjacent bins. The CTF for data symbols is then reconstructed by inverse transforms. Numerical results compare the proposed estimator with linear interpolation and MMSE, reporting near-MMSE performance at lower complexity and without requiring channel statistics.

Significance. The on-grid derivation from Eqs. (10) to (16) is clean and correctly identifies the scaling factors (N/d_t^2, M/d_f^2) and the support limits, providing a useful design guideline for pilot spacing in doubly selective channels. The method does not require channel statistical information and offers a complexity advantage over MMSE if the stated complexity holds. However, the claimed exactness for off-grid Doppler is not supported by the current analysis: Eq. (19) relies on a finite-N small-angle approximation, and Algorithm 1 depends on on-grid delays, a one-path-per-delay-bin assumption, and an externally supplied number of paths. These limitations are load-bearing for Theorem 2 and are not exercised by the simulations, which use on-grid delays and likely known P. With these caveats addressed, the approximate estimator is a plausible and practically interesting contribution.

major comments (3)
  1. [Section III-B, Eq. (19), Theorem 2] The 'exactly estimated' claim in Theorem 2 is not supported by the derivation. Step (a) in Eq. (19) replaces the ratio of sine functions by (k_i^F - k'_0 + k_0)/k_i^F, which is a small-angle approximation valid only for sufficiently large N, not an identity. Consequently, the estimated fractional Doppler \hat{k}_i^F in Eq. (20) carries an O(1/N) bias even in the noiseless limit, and the reconstructed CSF in Eq. (21) and CTF in Eq. (16) are not exactly equal to the true channel functions. The theorem should be restated as an approximate statement with an explicit error bound, or at minimum with a limit N → ∞ qualifier.
  2. [Section II-A and Theorem 2] Theorem 2 omits the on-grid delay assumption that is explicit in Section II-A, namely that l_i = M Δf τ_i is an integer. Algorithm 1 Step 4 estimates the delay as l_0, an integer bin, and Step 6 nulls the entire delay row; this is valid only when each path delay is on-grid and at most one path occupies each resolvable delay bin. As written, Theorem 2 claims exact estimation under conditions only on the ranges of ν_i and τ_i, which is misleading for practical wideband channels with continuous delays and multiple paths per delay bin. The theorem statement and Remark 4 must list all underlying assumptions and restrict the claims accordingly.
  3. [Section IV, Figs. 2 and 3] The numerical results do not test the regime in which Theorem 2 is claimed to hold. The captions of Figs. 2 and 3 specify on-grid delays, and the extended vehicular A channel is not used with its natural continuous delays. No experiment with off-grid delays or with multiple paths sharing one delay bin is reported, and the number of paths \hat{P} is presumably known or obtained by an unspecified energy threshold. The conclusion that the proposed design 'offers near-optimal performance' is therefore validated only under conditions that match the favorable assumptions, not under the general conditions implied by Theorem 2.
minor comments (6)
  1. [Algorithm 1, Step 6] In Step 6, the index set for k is written as S^1_M; since k is the Doppler index of length N, it should be S^1_N.
  2. [Remark 3] Remark 3 attributes the off-grid error to 'DD domain aliasing,' but the effect is spectral leakage of the sinc-like Dirichlet kernel; consider using 'leakage' for clarity.
  3. [Abstract and Section V] The claim that the method works 'without requiring prior information' is too strong: Algorithm 1 requires \hat{P} as an input and relies on the one-path-per-delay-bin structural assumption. The claim should be qualified.
  4. [Section III-B, Algorithm 1] Algorithm 1 is taken from prior work [12] with minimal adaptation; the specific choices for the energy threshold and the handling of multiple paths in the same delay bin are not described, which makes the method difficult to reproduce.
  5. [Section IV, complexity claim] The claimed complexity O(N log N + M log M + P) should be justified for Eq. (21) and Algorithm 1: the reconstruction in Eq. (21) as written is a sum over paths and, if evaluated directly for all (k,l), would cost more than the stated complexity.
  6. [Notation] The set notation S^α_β is described as the 'positive integer set from 0 to α−1'; since 0 is included, 'nonnegative integers' would be more accurate.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: the CSF estimate is derived from the Fourier-pair relationship with pilot sampling; Theorem 2's exact wording is a correctness overstatement, not a circular step.

full rationale

The claimed derivation chain is self-contained. Starting from the OFDM input-output model, the paper defines the pilot-sampled CTF and obtains the periodic CSF by a 2-D DFT/IDFT. The geometric-series expressions in Eqs. (12)-(14) are algebraic identities, and Theorem 1 follows from the support conditions plus the choice of scaling factors ~M = M/d_f^2 and ~N = N/d_t^2, which are solved from the requirement that one period equals the full-band DFT; no parameter is fitted to the simulated channel or to the MMSE benchmark. Algorithm 1 is taken from the external prior work [12] with no author overlap with the present paper, so the off-grid Doppler estimator is not a self-citation chain, and the paper's own contribution is the pilot-arrangement/compactness link. The only notable weakness is that Theorem 2 calls the off-grid case 'exactly estimated' although Eq. (19)(a) is a finite-N approximation and Remark 3 admits aliasing errors; that is an overstatement of correctness, not a circular reduction. Likewise, the on-grid delay assumption and the one-path-per-delay-bin premise of Algorithm 1 are modeling limitations, not circularity.

Assumptions & free parameters 2 free parameters · 6 assumptions · 0 invented entities

The central claim rests on standard Fourier sampling relations plus several domain assumptions about the channel. No new physical entities are introduced. The pilot spacings are hand-picked in the simulations, and the path-count threshold is an unspecified free parameter inherited from [12].

free parameters (2)
  • Pilot spacings dt and df = dt = 4, df = 4
    Chosen by hand in Section IV to satisfy the compactness conditions of Theorems 1 and 2; the method's validity is tied to these values, and larger spacings would violate the support bound.
  • Energy threshold for number of paths P = not specified
    Algorithm 1 requires P as input, said to be observable via an energy threshold in [12]; the paper does not define the threshold or analyze its sensitivity.
assumptions (6)
  • domain assumption All path delays are on-grid, so l_i = M * delta_f * tau_i is an integer
    Stated in Section II-A after Eq. (2), and used throughout the CSF model and Algorithm 1; continuous delays break the exact recovery claim.
  • domain assumption The ICI from Doppler is negligible, so the TF input-output relation is diagonal per subcarrier
    Eq. (3) approximates Eq. (2) under k_i << N; the paper justifies with N*T*nu_max = 2.06 << 64 in the numerical setup, but ICI is otherwise not modeled.
  • domain assumption CSF is quasi-static over the observation window
    Section III relies on the stationary-region property from [8] to justify estimating one CSF from a frame of pilots.
  • ad hoc to paper At most one path occupies each resolvable delay grid
    Stated before Algorithm 1; the peak-cancellation loop in Step 6 deletes an entire delay row, so two paths in one delay bin are not separable.
  • standard math N is large enough to justify the small-angle approximation in Eq. (19)
    The replacement sin(pi*dt*(k_Fi - k'_0 + k_0)/N) / sin(pi*dt*k_Fi/N) approximately equals (k_Fi - k'_0 + k_0)/k_Fi is used to obtain a closed-form fractional Doppler estimate.
  • domain assumption CSF support is contained in the claimed rectangle
    Theorems 1 and 2 require Doppler in [-1/(2*dt*T), 1/(2*dt*T) - 1/(N*T)] and delay in [0, 1/(df*delta_f) - 1]; the method is only exact within this support.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Channel Spreading Function-Inspired Channel Transfer Function Estimation for OFDM Systems with High-Mobility." pith.science (2026). https://pith.science/paper/2S54SQVP

@misc{pith2026241207074,
  author       = {Pith},
  title        = {Pith review of: Channel Spreading Function-Inspired Channel Transfer Function Estimation for OFDM Systems with High-Mobility},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/2S54SQVP}},
  note         = {Machine review of arXiv:2412.07074}
}
read the original abstract

In this letter, we propose a novel channel transfer function (CTF) estimation approach for orthogonal frequency division multiplexing (OFDM) systems in high-mobility scenarios, that leverages the stationary properties of the delay-Doppler domain channel spreading function (CSF). First, we develop a CSF estimation model for OFDM systems that relies solely on discrete pilot symbols in the time-frequency (TF) domain, positioned at predefined resource elements. We then present theorems to elucidate the relationship between CSF compactness and pilot spacing in the TF domain for accurate CSF acquisition. Based on the estimated CSF, we finally estimate the CTF for data symbols. Numerical results show that, in high-mobility scenarios, the proposed approach outperforms traditional interpolation-based methods and closely matches the optimal estimator in terms of estimation accuracy. This work may pave the way for CSF estimation in commercial OFDM systems, benefiting high-mobility communications, integrated sensing and communications, and related applications.

Figures

Figures reproduced from arXiv: 2412.07074 by the authors.

Figure 1
Figure 1. Lattice-type pilot arrangement for the OFDM system. [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. Performance of CTF estimation with on-grid delays and on-grid [PITH_FULL_IMAGE:figures/full_fig_p005_2.png] view at source ↗
Figure 3
Figure 3. Performance of CTF estimation with on-grid delays and off-grid [PITH_FULL_IMAGE:figures/full_fig_p005_3.png] view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

12 extracted references · 11 canonical work pages

  1. [9]

    Channel Estimation, Interpolation and Extrapolation in Doubly-dispersive Channels,

    Z. Gong, et al., “Channel Estimation, Interpolation and Extrapolation in Doubly-dispersive Channels,” arXiv:2408.09381 [SP], Aug. 2024

    arXiv 2024

  2. [10]

    Delay-Doppler channel estimation by leveraging the ambiguity function in OFDM systems,

    H. P. H. Shaw, J. Yuan and M. Rowshan, “Delay-Doppler channel estimation by leveraging the ambiguity function in OFDM systems,” in Proc. IEEE Int. Conf. Commun. (ICC) , Rome, Italy, 2023, pp. 307-313

    work page 2023

  3. [1]

    What will 5G be?

    J. G. Andrews, et al., “What will 5G be?” IEEE J. Sel. Areas Commun. , vol. 32, no. 6, pp. 1065–1082, Jun. 2014

    work page 2014

  4. [2]

    Y . S. Cho, J. Kim, W. Y . Yang, and C. G. Kang, MIMO-OFDM wireless communications with MATLAB . Hoboken, NJ, USA: Wiley, 2010

    work page 2010

  5. [3]

    Deep learning-based channel estimation,

    M. Soltani, et al., “Deep learning-based channel estimation,” IEEE Commun. Lett., vol. 23, no. 4, pp. 652–655, Apr. 2019

    work page 2019

  6. [4]

    Orthogonal delay-Doppler division multiplexing modulation,

    H. Lin and J. Yuan, “Orthogonal delay-Doppler division multiplexing modulation,” IEEE Trans. Wirel. Commun. , vol. 21, no. 12, pp. 11024- 11037, Dec. 2022

    work page 2022

  7. [5]

    On the coexistence of OTFS with OFDM-based wireless communication systems,

    A. Shafie, et al., “On the coexistence of OTFS with OFDM-based wireless communication systems,” IEEE Trans. Commun. , vol. 01, no. 01, pp. 1-1, June 2024

    work page 2024

  8. [6]

    Interference cancellation and iterative detection for orthogonal time frequency space modulation,

    P. Raviteja et al., “Interference cancellation and iterative detection for orthogonal time frequency space modulation,” IEEE Trans. Wirel. Commun., vol. 17, no. 10, pp. 6501–15, Oct. 2018

    work page 2018

Show all 12 references
  1. [7]

    OTFCS-modulated waveform design for joint grant-free random access and positioning in C-V2X,

    Y . Ma et al., “OTFCS-modulated waveform design for joint grant-free random access and positioning in C-V2X,” IEEE J. Sel. Areas Commun., vol. 42, no. 1, pp. 103-119, Jan. 2024

    2024

  2. [8]

    Hlawatsch and G

    F. Hlawatsch and G. Matz, Wireless communications over rapidly time- varying channels, Academic press, 2011

    2011

  3. [11]

    On the effective- ness of OTFS for joint radar parameter estimation and communication,

    L. Gaudio, M. Kobayashi, G. Caire, and G. Colavolpe, “On the effective- ness of OTFS for joint radar parameter estimation and communication,” IEEE Trans. Wirel. Commun., vol. 19, no. 9, pp. 5951–5965, Sep. 2020

    2020

  4. [12]

    Deterministic pilot design and channel estimation for downlink massive MIMO-OTFS systems in presence of the fractional doppler,

    D. Shi, W. Wang, L. You, X. Song, Y . Hong, X. Gao, and G. Fettweis, “Deterministic pilot design and channel estimation for downlink massive MIMO-OTFS systems in presence of the fractional doppler,” IEEE Trans. Wirel. Commun., vol. 20, no. 11, pp. 7151-7165, 2021

    2021

Pith tools

Reviewed August 11, 2026 · model on record in the stance chip above.