Pith. sign in

REVIEW 4 major objections 5 minor 1 cited by

Superimposed DMRS for Spectrally Efficient 6G Uplink Multi-User OFDM: Classical vs AI/ML Receivers

T0 review · 4 major / 5 minor · reviewed 2026-08-06 · deepseek-v4-flash

Pith's one-line read A convolutional receiver can estimate channels from superimposed pilots and beat orthogonal pilots at high SNR.

desk verdict A solid simulation comparison with a useful result, but the 'consistently outperforms' claim is contradicted by the paper's own 4-UE QPSK scenario, and the OCC orthogonality argument ignores channel selectivity. read the letter →

arxiv 2506.20248 v1 pith:NTAPEJGB submitted 2025-06-25 eess.SP

classification eess.SP
keywords superimposedDMRSorthogonalcovercodesDeepRxneuralreceiveruplinkMU-MIMOOFDMchannelestimation6G
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 argues that a 6G uplink can send data on every time-frequency resource element by superimposing low-power reference signals on top of data, rather than reserving separate pilot resources, and that a learned convolutional receiver can make that scheme pay off. The authors configure the superimposed DMRS with orthogonal cover codes so the pilot components of different users remain separable, then compare a DeepRx neural receiver against a classical iterative receiver that cancels data-to-pilot interference. In simulated single-user and multi-user MIMO uplinks, DeepRx outperforms the classical receiver in every superimposed-DMRS scenario tested, and at high signal-to-noise ratio the superimposed scheme with DeepRx delivers higher throughput than orthogonal 5G-style DMRS. If the result carries into practice, the pilot overhead that currently grows with the number of users and layers could be largely eliminated.

What carries the argument

The carrying mechanism is a code-domain pilot plan combined with a hybrid classical/neural receiver. Each user or layer is assigned an orthogonal cover code (OCC) sequence spread over the entire time-frequency grid; because the inner product of two OCC sequences is zero, the superimposed pilot components are nominally orthogonal even though data symbols occupy the same resource elements. DeepRx starts from a least-squares channel estimate, refines it with a convolutional ResNet, equalizes through parallel LMMSE-type and maximum-ratio-combining branches, and maps the refined symbol estimates to bit log-likelihood ratios with a second ResNet, training the whole pipeline end-to-end with a binary cross-entropy loss.

What would settle it

Re-run the DeepRx-versus-classical comparison on a channel with a large delay spread, or with shorter OCC groups, so the channel coefficient changes inside each OCC group; if the SI DMRS throughput advantage over orthogonal DMRS disappears or reverses, the OCC separability assumption is the cause.

Watch

Extended reading notes

Core claim

The paper's central claim is that a CNN-based receiver trained end-to-end can estimate the channel from superimposed DMRS and detect data reliably enough that the superposition becomes a net throughput win. DeepRx beats the classical iterative receiver in all tested SI DMRS configurations, and SI DMRS with DeepRx provides higher throughput than orthogonal DMRS at high SNR across single-stream, two-stream, two-user, and four-user uplink cases. The authors attribute the gain to DeepRx learning to treat residual data-to-pilot interference as structured information rather than noise, while the classical receiver's channel estimates degrade under higher-order modulation.

Load-bearing premise

The scheme assumes the channel is flat enough across each OCC group that the orthogonal cover codes keep the superimposed pilots separable; if the channel response varies within a group, inter-layer pilot interference returns and the reported gains may shrink.

Editorial extensions

If this is right

  • At high signal-to-noise ratio, SI DMRS with DeepRx yields higher throughput than orthogonal DMRS in the tested configurations, so removing dedicated pilot resources can translate directly into spectral efficiency gains when channel conditions are good.
  • DeepRx outperforms the classical iterative receiver in every SI DMRS scenario tested, indicating that learned receivers handle data-to-pilot interference better than hand-designed cancellation loops.
  • At low SNR, orthogonal DMRS remains the better choice even with DeepRx, so a practical system would likely switch between the two schemes based on operating point.
  • The classical iterative receiver with SI DMRS is competitive only under QPSK; with 16-QAM and 64-QAM its channel estimates are too imprecise, so the viability of superimposed pilots depends on having a sophisticated receiver.
  • A shallower DeepRx variant with fewer ResNet blocks reaches nearly the same throughput, so the reported gains do not require the largest model.

Reading between the lines

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

  • Because the paper's orthogonality argument uses the inner product of the cover codes without accounting for the channel response multiplying each OCC symbol, the gains should be re-tested in more frequency-selective channels or with shorter OCC groups; if the channel varies inside a group, inter-layer pilot interference could return.
  • The same superposition could be combined with per-link adaptation: choose orthogonal DMRS at low SNR, SI DMRS at high SNR, and optimize the pilot-to-data power ratio per modulation order rather than fixing it per scenario.
  • The conclusion's observation that control channels occupy the first OFDM symbols suggests SI DMRS would be especially valuable when few symbols remain for data; a system-level study could quantify this benefit.
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.

Referee Report

4 major / 5 minor

Summary. This paper studies superimposed DMRS (SI-DMRS) transmission for uplink OFDM in single- and multi-user MIMO. The authors propose an enhanced DeepRx CNN receiver that takes SI-DMRS signals and pilot configurations as input, uses an LS channel estimate and a Channel-estimation CNN, then performs equalization and demapping through Detector and Demapper CNNs trained end-to-end with a BCE loss. They also design a classical LS-based iterative receiver with sliding-window smoothing and LDPC decoding in the loop. The two receiver families are compared against orthogonal 5G DMRS and against genie-aided linear detection in Sionna simulations with UMa channels, 1/2/4 users, 1/2 transmit antennas, 4/16 receive antennas, and QPSK/16QAM/64QAM. The central claims are that DeepRx consistently outperforms the classical receivers with SI-DMRS and that DeepRx with SI-DMRS yields throughput gains over DeepRx with orthogonal DMRS at high SNR.

Significance. If the reported results hold, the paper would provide a useful data point for the 6G discussion on AI/ML-based receivers and superimposed reference signals, showing that a neural receiver can remove pilot overhead while retaining reliable channel estimation. The main strengths are the reproducible simulation setup built on Sionna, the detailed reporting of architecture and scenario parameters in Tables I-III, and the straightforward LS/iterative formulation in Section IV, which is internally consistent. The paper also reports a depth ablation (Figure 9). However, the empirical comparison currently overclaims relative to its own results, and the OCC-based orthogonality argument and tuning fairness need to be resolved before the central comparison can be considered established.

major comments (4)
  1. [Abstract and Section I vs. Section V-B2] The abstract and introduction claim that DeepRx 'consistently outperforms' the classical receivers in all SI-DMRS scenarios, but Section V-B2 reports that 'the only scenario where the conventional system achieves higher performance with SI DMRS is the case with QPSK and 4 UEs' (Figure 8, top). Because the authors attribute this outcome to the robustness of QPSK, it appears to be a systematic regime rather than a statistical fluctuation. The headline claim should be weakened to 'in the majority of tested scenarios,' or the QPSK/4-UE scenario should be explicitly excluded from the stated scope.
  2. [Section III] The orthogonality argument for OCC-based SI DMRS is incomplete. The statement that 'the inner product of two OCC sequences is zero' ignores the channel response: the received contribution of layer ell at RE (i,j) is H_{i,j}^{(ell)} p_{i,j}^{(ell)}, so after correlation with p^{(k)} the residual inter-layer term is sum_{i,j} H_{i,j}^{(ell)} p_{i,j}^{(ell)} (p_{i,j}^{(k)})^*, which is not zero unless the channel is constant over the OCC spreading block. Since the OCCs are spread over the entire time-frequency grid and the channel is frequency-selective (UMa, 3.5 GHz, 72 subcarriers), the claimed 'nearly eliminated' inter-layer pilot interference is not justified. Please specify the OCC group size relative to the channel coherence block or numerically evaluate the residual cross-layer interference.
  3. [Section V-A and Table II] The comparison is asymmetric in tuning effort. The text states that 'to achieve optimal performance from the iterative receiver, we ... conducted an extensive search for the optimal SI DMRS power ratio E, number of iterations, and sliding window sizes,' while for DeepRx no equivalent search is reported; the DeepRx power ratios in Table II appear fixed per scenario, and the only DeepRx hyperparameter studied is depth (Section V-C). If the classical baseline is individually optimized and the ML receiver is not, the conclusion that 'DeepRx consistently outperforms' may partly reflect unequal optimization. Please state how the DeepRx E values were selected and provide a sensitivity analysis over E (and ideally over window-size inputs to the channel-estimation CNN) for DeepRx.
  4. [Section V and Figures 5-8] All throughput and BER curves are reported without error bars, random seeds, or the number of independent channel realizations used for evaluation. The paper states that training samples were generated on-the-fly, but this does not quantify evaluation uncertainty. Given that the central claim is a comparative performance statement, confidence intervals or multiple-seed results are needed to establish that the observed gaps are not due to finite-sample noise, particularly in the QPSK/4-UE case where the conventional receiver is reported to win.
minor comments (5)
  1. [Table II] Several window-size tuples in Table II are missing closing brackets (e.g., '(2,14]' in the 64-QAM rows); please correct these typographical errors.
  2. [Section IV-A2] The definition of v_{i,j}^{(k,u-1)} in Eq. (9) is hard to parse because the same symbol is used for a sum over all users and then a second sum over k' not equal to k; consider introducing separate names for these two interference terms.
  3. [Section V-B2] The explanation that the QPSK/4-UE exception is 'most likely due to the high robustness of QPSK' is speculative; please replace it with a quantitative argument, such as the ratio of the residual pilot-interference floor to the decision-region size.
  4. [Section V-B1] The observation that DeepRx with orthogonal DMRS can outperform a linear genie-aided receiver with perfect channel knowledge is surprising and would benefit from an explicit explanation of the nonlinear detection gain earlier in the section, not only in the final paragraph.
  5. [Section VI] The conclusion claims 'keeping inference complexity within practical bounds,' but no complexity measurements are reported; either add runtime or FLOPs comparisons or soften this statement.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: the DeepRx versus classical receiver comparison is an independent simulation result; self-citations are method references, not load-bearing evidence.

full rationale

The paper's central claims are empirical simulation comparisons produced by Sionna link-level simulations, not quantities fitted from the data they are used to predict. DeepRx is an existing neural receiver architecture whose training and block design are cited from prior work [8], but the paper retrains dedicated models for SI and orthogonal DMRS and evaluates them against newly designed classical iterative receivers; the performance ordering is an outcome of the simulation, not an input. The SI-DMRS power ratios and sliding-window sizes in Table II are operational parameters tuned for the comparison, and the paper explicitly states that the iterative receiver was extensively tuned; these choices affect the comparison but do not define the conclusion "DeepRx outperforms" by construction. The OCC orthogonality statement in Section III ('the inner product of two OCC sequences is zero') is a mathematical property of the pilot sequences; whether it survives frequency-selective channels is a modeling-validity concern, not a circular step. Likewise, Section V-B2's admission that the conventional receiver wins with QPSK and 4 UEs contradicts the abstract's 'consistently outperforms' phrasing, but this is an internal-consistency and correctness issue, not circularity. Self-citations to DeepRx papers are references to the method being extended and to training details; they do not supply the new result. No load-bearing argument reduces to its own inputs by definition.

Assumptions & free parameters 3 free parameters · 3 assumptions · 0 invented entities

No new physical entities are introduced; the OCC-based superimposed DMRS is a transmission scheme, not a new entity. The free parameters are operational settings (pilot power ratios, smoothing windows) that are tuned per scenario, and the axioms are standard simulation-modeling assumptions plus the OCC orthogonality premise.

free parameters (3)
  • SI DMRS pilot power ratio E (iterative receiver) = 0.14, 0.22, 0.30, 0.22, 0.35, 0.43, 0.35, 0.43, 0.24, 0.55 (per scenario, Table II)
    Chosen via extensive search per scenario (Section V-A); directly controls the pilot-data trade-off and the feasibility of channel estimation for the classical baseline.
  • SI DMRS pilot power ratio E (DeepRx) = 0.035, 0.07, 0.14 (per scenario, Table II)
    Hand-selected for the neural receiver; no search procedure reported, and the values are smaller than those for the iterative receiver, affecting the comparison.
  • Sliding window sizes (W_f, W_t) and iteration count U (iterative receiver) = e.g., [(12,14), (6,14)], [(8,14), (6,14), (6,14), (4,14)] per scenario, Table II
    Selected via extensive search; the number of windows equals the number of iterations U and determines smoothing severity, which is critical for the classical receiver's performance.
assumptions (3)
  • domain assumption OCC sequences assigned to different layers/users remain orthogonal in the received signal over the time-frequency grid.
    Section III: 'the inner product of two OCC sequences is zero'. This ignores that the frequency-selective channel multiplies the OCC per RE, so orthogonality is only approximate when the channel varies across the OCC group.
  • domain assumption The received signal model y = H x + n with AWGN and perfect synchronization is valid.
    Eq. (1) is the basis for all estimators; no CFO, timing offset, or hardware impairments are modeled.
  • domain assumption The UMa channel model implemented in the frequency domain (time-invariant within the slot) is representative of 6G uplink conditions.
    Table I specifies 'Channel implementation frequency domain' and UMa at 3.5 GHz; the reported gains depend on this model.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Superimposed DMRS for Spectrally Efficient 6G Uplink Multi-User OFDM: Classical vs AI/ML Receivers." pith.science (2026). https://pith.science/paper/NTAPEJGB

@misc{pith2026250620248,
  author       = {Pith},
  title        = {Pith review of: Superimposed DMRS for Spectrally Efficient 6G Uplink Multi-User OFDM: Classical vs AI/ML Receivers},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/NTAPEJGB}},
  note         = {Machine review of arXiv:2506.20248}
}
read the original abstract

Fifth-generation (5G) systems utilize orthogonal demodulation reference signals (DMRS) to enable channel estimation at the receiver. These orthogonal DMRS-also referred to as pilots-are effective in avoiding pilot contamination and interference from both the user's own data and that of others. However, this approach incurs a significant overhead, as a substantial portion of the time-frequency resources must be reserved for pilot transmission. Moreover, the overhead increases with the number of users and transmission layers. To address these limitations in the context of emerging sixth-generation (6G) systems and to support data transmission across the entire time-frequency grid, the superposition of data and DMRS symbols has been explored as an alternative DMRS transmission strategy. In this study, we propose an enhanced version of DeepRx, a deep convolutional neural network (CNN)-based receiver, capable of estimating the channel from received superimposed (SI) DMRS symbols and reliably detecting the transmitted data. We also design a conventional receiver for comparison, which estimates the channel from SI DMRS using classical signal processing techniques. Extensive evaluations in both uplink single-user and multi-user scenarios demonstrate that DeepRx consistently outperforms the conventional receivers in terms of performance.

Figures

Figures reproduced from arXiv: 2506.20248 by the authors.

Figure 1
Figure 1. Block diagram of the considered UL MU-MIMO OFDM system. [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. DMRS transmission schemes considered in this paper. [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 3
Figure 3. An example of SI DMRS using OCC configuration for four transmission layers. Here, time- and frequency-domain OCCs are configured to ensure [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗
Figures from the paper (6 more)
Figure 4
Figure 4. Figure 4: A high-level depiction of the DeepRx architecture, used both for superimposed and orthogonal DMRS. [PITH_FULL_IMAGE:figures/full_fig_p005_4.png]
Figure 5
Figure 5. Figure 5: Coded BER and throughput vs. SNR for SU MIMO 1x4 QPSK (top) 16QAM (middle) and 64QAM (bottom). [PITH_FULL_IMAGE:figures/full_fig_p008_5.png]
Figure 6
Figure 6. Figure 6: Coded BER and throughput vs. SNR for SU MIMO 2x4 QPSK (top), 16QAM (middle), and 64QAM (bottom). [PITH_FULL_IMAGE:figures/full_fig_p009_6.png]
Figure 7
Figure 7. Figure 7: Coded BER and throughput vs. SNR for MU MIMO, 2x(1x4) 16QAM (top) and 64QAM (bottom). [PITH_FULL_IMAGE:figures/full_fig_p010_7.png]
Figure 8
Figure 8. Figure 8: Coded BER and throughput vs. SNR for MU MIMO, 4x(1x16) QPSK (top) and 64QAM (bottom). [PITH_FULL_IMAGE:figures/full_fig_p011_8.png]
Figure 9
Figure 9. Figure 9: Throughput results with different depths (number of ResNet blocks) [PITH_FULL_IMAGE:figures/full_fig_p011_9.png]

Discussion (0). Continue with ORCID 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

  1. Learning-Aided Iterative Receiver for Superimposed Pilots: Design and Experimental Evaluation

    cs.IT 2025-07 conditional novelty 5.0 of 10

    A superimposed-pilot MIMO-OFDM receiver with variational and deep-learning channel estimators improves throughput and block error rate over orthogonal-pilot baselines in simulations and over-the-air tests.

Reference graph

Works this paper leans on

25 extracted references · 24 canonical work pages · cited by 1 Pith paper

  1. [1]

    An introduction to deep learning for the physical layer,

    T. O’Shea and J. Hoydis, “An introduction to deep learning for the physical layer,”IEEE Transactions on Cognitive Communications and Networking, vol. 3, no. 4, pp. 563–575, 2017

  2. [2]

    3GPP TR 38.843 – Study on Artificial Intelligence (AI)/Machine Learning (ML) for NR air interface,

    3GPP TR-38.843, “3GPP TR 38.843 – Study on Artificial Intelligence (AI)/Machine Learning (ML) for NR air interface,” June 2022. 12

  3. [3]

    3GPP TR 38.744 V1.0.0: Study on Artificial Intelligence (AI)/Machine Learning (ML) for mobility in NR,

    3rd Generation Partnership Project (3GPP), “3GPP TR 38.744 V1.0.0: Study on Artificial Intelligence (AI)/Machine Learning (ML) for mobility in NR,” 3rd Generation Partnership Project (3GPP), Technical Report TR 38.744, June 2025. [Online]. Available: https://www.3gpp.org/ftp/Specs/archive/38 series/38.744/38744-100.zip

  4. [4]

    Machine Learning-enhanced Receive Processing for MU-MIMO OFDM Sys- tems,

    M. Goutay, F. A. Aoudia, J. Hoydis, and J.-M. Gorce, “Machine Learning-enhanced Receive Processing for MU-MIMO OFDM Sys- tems,” in2021 IEEE 22nd International Workshop on Signal Processing Advances in Wireless Communications (SPAWC), 2021, pp. 246–250

  5. [5]

    DeepRx MIMO: Convolutional MIMO Detection with Learned Multiplicative Transformations,

    D. Korpi, M. Honkala, J. M. Huttunen, and V . Starck, “DeepRx MIMO: Convolutional MIMO Detection with Learned Multiplicative Transformations,” inICC 2021 - IEEE International Conference on Communications, 2021, pp. 1–7

  6. [6]

    Adaptive NN-based OFDM Receivers: Computational Complexity vs. Achievable Performance,

    M. B. Fischer, S. D ¨orner, F. Krieg, S. Cammerer, and S. t. Brink, “Adaptive NN-based OFDM Receivers: Computational Complexity vs. Achievable Performance,” in2022 56th Asilomar Conference on Signals, Systems, and Computers, 2022, pp. 194–199

  7. [7]

    iDeepRx Enabled 100 Gb/s DFT-s-OFDM Data Transmission Over 220 GHz Testbed,

    W. Qi, C. Ye, D. J. Korpi, C. Gong, Y . Jin, and T. Yang, “iDeepRx Enabled 100 Gb/s DFT-s-OFDM Data Transmission Over 220 GHz Testbed,” inICC 2023 - IEEE International Conference on Commu- nications, 2023, pp. 3684–3689

  8. [8]

    DeepRx: Fully convolu- tional deep learning receiver,

    M. Honkala, D. Korpi, and J. M. J. Huttunen, “DeepRx: Fully convolu- tional deep learning receiver,”IEEE Transactions on Wireless Commu- nications, vol. 20, no. 6, pp. 3925–3940, 2021

Show all 25 references
  1. [9]

    Hybrid- deeprx: Deep learning receiver for high-evm signals,

    J. Pihlajasalo, D. Korpi, M. Honkala, J. M. J. Huttunen, T. Riihonen, J. Talvitie, A. Brihuega, M. A. Uusitalo, and M. Valkama, “Hybrid- deeprx: Deep learning receiver for high-evm signals,” in2021 IEEE 32nd Annual International Symposium on Personal, Indoor and Mobile Radio C...

  2. [10]

    A Neural Receiver for 5G NR Multi-User MIMO,

    S. Cammerer, F. A. Aoudia, J. Hoydis, A. Oeldemann, A. Roessler, T. Mayer, and A. Keller, “A Neural Receiver for 5G NR Multi-User MIMO,” in2023 IEEE Globecom Workshops (GC Wkshps), 2023, pp. 329–334

  3. [11]

    End-to-End Learning for OFDM: From Neural Receivers to Pilotless Communication,

    F. Ait Aoudia and J. Hoydis, “End-to-End Learning for OFDM: From Neural Receivers to Pilotless Communication,”IEEE Transactions on Wireless Communications, vol. 21, no. 2, pp. 1049–1063, 2022

  4. [12]

    Waveform Learning for Next-Generation Wireless Communica- tion Systems,

    ——, “Waveform Learning for Next-Generation Wireless Communica- tion Systems,”IEEE Transactions on Communications, vol. 70, no. 6, pp. 3804–3817, 2022

  5. [13]

    End-to-End Learning of OFDM Waveforms with PAPR and ACLR Constraints,

    M. Goutay, F. A. Aoudia, J. Hoydis, and J.-M. Gorce, “End-to-End Learning of OFDM Waveforms with PAPR and ACLR Constraints,” in 2021 IEEE Globecom Workshops (GC Wkshps), 2021, pp. 1–6

  6. [14]

    Deep learning-based pilotless spatial multiplexing,

    D. Korpi, M. Honkala, and J. M. Huttunen, “Deep learning-based pilotless spatial multiplexing,” in2023 57th Asilomar Conference on Signals, Systems, and Computers, 2023, pp. 1025–1029

  7. [15]

    End-to-end waveform learning through joint optimization of pulse and constellation shaping,

    F. A. Aoudia and J. Hoydis, “End-to-end waveform learning through joint optimization of pulse and constellation shaping,” in2021 IEEE Globecom Workshops (GC Wkshps), 2021, pp. 1–6

  8. [16]

    End-to-End Learning for Uplink MU- SIMO Joint Transmitter and Non-Coherent Receiver Design in Fading Channels,

    S. Xue, Y . Ma, and N. Yi, “End-to-End Learning for Uplink MU- SIMO Joint Transmitter and Non-Coherent Receiver Design in Fading Channels,”IEEE Transactions on Wireless Communications, vol. 20, no. 9, pp. 5531–5542, 2021

  9. [17]

    Partial-Data Superimposed Training With Data Precoding for OFDM Systems,

    J. C. Estrada-Jim ´enez and M. J. Fern ´andez-Getino Garc´ıa, “Partial-Data Superimposed Training With Data Precoding for OFDM Systems,”IEEE Transactions on Broadcasting, vol. 65, no. 2, pp. 234–244, 2019

  10. [18]

    Superimposed Versus Regular Pilots for Hardware Impaired Rician- Faded Cell-Free Massive MIMO Systems,

    H. Haritha, D. N. Amudala, R. Budhiraja, and A. K. Chaturvedi, “Superimposed Versus Regular Pilots for Hardware Impaired Rician- Faded Cell-Free Massive MIMO Systems,”IEEE Transactions on Com- munications, pp. 1–1, 2024

  11. [19]

    Superimposed Pilot Transmission in Cell-Free Massive MIMO With Non-Ideal RF Responses,

    Y . Zhang, H. Zhao, W. Xia, F. Gao, L. Yang, and H. Zhu, “Superimposed Pilot Transmission in Cell-Free Massive MIMO With Non-Ideal RF Responses,”IEEE Transactions on Vehicular Technology, vol. 71, no. 12, pp. 12 856–12 868, 2022

  12. [20]

    AI-Driven Iterative Receiver for Superimposed Pilot Schemes in MIMO-OFDM Systems,

    X. Li, X. Zhou, J. Zhang, C.-K. Wen, and S. Jin, “AI-Driven Iterative Receiver for Superimposed Pilot Schemes in MIMO-OFDM Systems,” in2025 IEEE Wireless Communications and Networking Conference (WCNC), 2025, pp. 1–6

  13. [21]

    Interference cancellation based neural receiver for superimposed pilot in multi-layer transmission,

    X. Han, T. Wenqiang, J. Shi, L. Wendong, S. Jia, S. Zhihua, and Z. Zhi, “Interference cancellation based neural receiver for superimposed pilot in multi-layer transmission,”China Communications, vol. 22, no. 1, pp. 75–88, 2025

  14. [22]

    Trimming the Fat from OFDM: Pilot- and CP-less Communication with End-to-end Learning,

    F. A. Aoudia and J. Hoydis, “Trimming the Fat from OFDM: Pilot- and CP-less Communication with End-to-end Learning,” in2021 IEEE Inter- national Conference on Communications Workshops (ICC Workshops), 2021, pp. 1–6

  15. [23]

    Deep Learn- ing OFDM Receivers for Improved Power Efficiency and Coverage,

    J. Pihlajasalo, D. Korpi, M. Honkala, J. M. J. Huttunen, T. Riihonen, J. Talvitie, A. Brihuega, M. A. Uusitalo, and M. Valkama, “Deep Learn- ing OFDM Receivers for Improved Power Efficiency and Coverage,” IEEE Transactions on Wireless Communications, vol. 22, no. 8, pp. 5518–5...

  16. [24]

    Detection of Impaired OFDM Waveforms Using Deep Learning Receiver,

    J. Pihlajasalo, D. Korpi, T. Riihonen, J. Talvitie, M. A. Uusitalo, and M. Valkama, “Detection of Impaired OFDM Waveforms Using Deep Learning Receiver,” in2022 IEEE 23rd International Workshop on Signal Processing Advances in Wireless Communication (SPAWC), 2022, pp. 1–5

  17. [25]

    Sionna: An Open-Source Library for Next-Generation Physical Layer Research,

    J. Hoydis, S. Cammerer, F. Ait Aoudia, A. Vem, N. Binder, G. Marcus, and A. Keller, “Sionna: An Open-Source Library for Next-Generation Physical Layer Research,”arXiv preprint, Mar. 2022. SAJAD REZAIEreceived his B.Sc. and M.Sc. degrees in Electrical Engineering from Amirkabir...

Pith tools

Reviewed August 6, 2026 · model on record in the stance chip above.