pith. machine review for the scientific record. sign in

arxiv: 2604.13500 · v1 · submitted 2026-04-15 · 💻 cs.NI

Recognition: unknown

Autoencoder-Based CSI Compression for Beyond Wi-Fi 8 Coordinated Beamforming

Ibrahim Aboushehada , Boris Bellalta , Giovanni Geraci , Lorenzo Galati Giordano
Authors on Pith no claims yet

Pith reviewed 2026-05-10 12:34 UTC · model grok-4.3

classification 💻 cs.NI
keywords CSI compressionautoencodercoordinated beamformingWi-FiIEEE 802.11bnchannel soundingmulti-AP coordinationMAPC
0
0 comments X

The pith

Autoencoder-based CSI compression reduces Wi-Fi sounding overhead by more than 50 percent and enables coordinated beamforming to outperform legacy transmissions.

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

This paper proposes an autoencoder to compress channel state information feedback in dense Wi-Fi networks using multi-access-point coordination. The aim is to overcome the high channel sounding overhead that currently limits coordinated beamforming performance under IEEE 802.11 standards. Evaluation in an event-driven simulator with realistic channels shows that the approach cuts overhead substantially compared to standard CSI compression. A compression ratio of one quarter offers the best tradeoff for accuracy and feedback size, resulting in lower data latency. The method allows coordinated beamforming to achieve higher throughput and better latency than legacy non-coordinated transmissions.

Core claim

The paper claims that integrating an autoencoder-based CSI compression into an IEEE 802.11bn-aligned MAC design reduces channel sounding overhead by more than 50 percent versus IEEE 802.11 CSI compression. A 1/4 compression ratio yields the optimal accuracy-to-feedback-size tradeoff and the lowest data latency. This enables coordinated beamforming to deliver substantial throughput gains and latency reductions over legacy transmissions without coordination, which otherwise suffer from excessive overhead that can make them underperform legacy in some cases.

What carries the argument

The autoencoder for CSI compression, which encodes full channel information into a compact representation for feedback and reconstructs it at the access points to support coordinated beamforming.

Load-bearing premise

Channels generated by Sionna RT will produce an autoencoder whose reconstruction accuracy holds up under real hardware impairments, mobility, and varying environments.

What would settle it

Deploy the trained autoencoder on a hardware testbed with live channel measurements and compare the resulting beamforming throughput and latency to the simulated results.

Figures

Figures reproduced from arXiv: 2604.13500 by Boris Bellalta, Giovanni Geraci, Ibrahim Aboushehada, Lorenzo Galati Giordano.

Figure 1
Figure 1. Figure 1: TXOP signaling structure example with 2 STAs for Co-BF with IEEE 802.11bn joint NDP channel sounding procedure. [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: Proposed autoencoder architecture for CSI compression with encoder layers (blue) decoder layers (green) and entropy bottleneck [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: Architecture of the channel attention module [14]. [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗
Figure 5
Figure 5. Figure 5: Violin plots of Co-BF channel sounding overhead at high [PITH_FULL_IMAGE:figures/full_fig_p005_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: 99th percentile and median latency with 4 STAs per AP. below, allowing sufficient time for A-MPDU transmission after channel sounding [PITH_FULL_IMAGE:figures/full_fig_p005_6.png] view at source ↗
Figure 7
Figure 7. Figure 7: Average throughput per STA comparison [PITH_FULL_IMAGE:figures/full_fig_p006_7.png] view at source ↗
Figure 8
Figure 8. Figure 8: 99th percentile and median data latency comparison. MHz case using either CSI compression technique. There are marginal gains in throughput and data latency with Co-BFAE compared to Co-BFST, but the performance is already well improved with Co-BF at this STA density. With 4 STAs per AP, Co-BF offers marginal latency reduc￾tions but with decent throughput gains compared to the legacy, which are further incr… view at source ↗
read the original abstract

Coordinated beamforming (Co-BF) is a key multi-access-point coordination (MAPC) technique for dense Wi-Fi deployments, but its performance can be hindered by the large channel state information (CSI) feedback required through channel sounding across overlapping basic service sets (OBSS). This work proposes an autoencoder (AE)-based CSI compression mechanism integrated into a standards-aligned IEEE 802.11bn MAC design. Using an event-driven simulator with realistic channels generated through Sionna RT, we evaluate the tradeoff between AE reconstruction accuracy and feedback size by measuring their impact on channel sounding overhead and data latency. Our results show that AE-based compression reduces channel sounding overhead by more than 50% compared to IEEE 802.11 CSI compression, with a compression ratio of 1/4 providing the best accuracy/feedback-size tradeoff for lowest data latency. Compared to legacy transmissions without MAPC, IEEE 802.11 CSI compression limits Co-BF due to high channel sounding overhead, causing it to underperform the legacy in some situations. However, AE-based CSI compression enables better Co-BF performance with substantial gains in throughput and data latency compared to legacy, demonstrating its promise as an enabler of efficient MAPC operation in future Wi-Fi systems.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

2 major / 2 minor

Summary. The paper proposes an autoencoder (AE)-based CSI compression scheme integrated into a standards-aligned IEEE 802.11bn MAC design for coordinated beamforming (Co-BF) in dense Wi-Fi deployments. Using an event-driven simulator fed by Sionna RT ray-traced channels, it evaluates the tradeoff between AE reconstruction accuracy and feedback size, claiming that AE compression reduces channel sounding overhead by more than 50% versus IEEE 802.11 CSI compression, with a 1/4 compression ratio providing the optimal accuracy/feedback-size tradeoff for minimal data latency and enabling substantial Co-BF throughput gains over legacy transmissions without MAPC.

Significance. If the reported overhead reductions and latency benefits hold, the work would be significant for enabling practical MAPC in future Wi-Fi by addressing the CSI feedback bottleneck in OBSS scenarios. Strengths include the standards-aligned MAC integration, use of an independent ray-tracing simulator for channel generation, and direct measurement of impacts on sounding overhead and data latency using standard metrics. These provide concrete, falsifiable predictions on compression ratios and performance tradeoffs.

major comments (2)
  1. [Evaluation] The central claims of >50% overhead reduction and superior Co-BF throughput rest on AE reconstruction accuracy measured exclusively in Sionna RT simulations (Evaluation section). No hardware impairments, mobility-induced Doppler, or environment-specific effects outside the ray-tracer are modeled, so if NMSE degrades under real conditions the measured gains versus 802.11 CSI compression will not materialize.
  2. [Results] The optimality of the 1/4 compression ratio for accuracy/feedback-size tradeoff is asserted without error-bar analysis on AE training runs or sensitivity tests to channel variations (Results section). This leaves the latency and throughput comparisons only partially verified, as small changes in reconstruction quality could alter the reported best tradeoff.
minor comments (2)
  1. [Proposed Method] Clarify the exact definition and training procedure for the autoencoder (e.g., loss function, network architecture details) in the Proposed Method section to allow reproducibility.
  2. [Abstract] The abstract states 'Beyond Wi-Fi 8' while the body refers to 802.11bn; ensure consistent terminology throughout.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the constructive and detailed comments. We address each major comment point by point below, indicating where revisions have been made to strengthen the manuscript while maintaining the integrity of our simulation-based evaluation.

read point-by-point responses

  1. Referee: [Evaluation] The central claims of >50% overhead reduction and superior Co-BF throughput rest on AE reconstruction accuracy measured exclusively in Sionna RT simulations (Evaluation section). No hardware impairments, mobility-induced Doppler, or environment-specific effects outside the ray-tracer are modeled, so if NMSE degrades under real conditions the measured gains versus 802.11 CSI compression will not materialize.

    Authors: We acknowledge that the evaluation relies on Sionna RT ray-tracing simulations, which do not incorporate hardware impairments, Doppler shifts from mobility, or additional environment-specific effects beyond the modeled scenarios. This is a recognized limitation of the current simulation study, as is common in initial assessments of new PHY/MAC techniques. Sionna RT has been validated against real measurements in the literature for indoor propagation, allowing us to isolate the impact of CSI compression on overhead and latency under controlled yet realistic channels. In the revised manuscript, we have added an explicit limitations paragraph in the Evaluation section discussing these factors and outlining future hardware testbed validation. We maintain that the reported overhead reductions and Co-BF gains remain valid within the simulation framework and provide a strong basis for further investigation. revision: partial

  2. Referee: [Results] The optimality of the 1/4 compression ratio for accuracy/feedback-size tradeoff is asserted without error-bar analysis on AE training runs or sensitivity tests to channel variations (Results section). This leaves the latency and throughput comparisons only partially verified, as small changes in reconstruction quality could alter the reported best tradeoff.

    Authors: We agree that statistical robustness would strengthen the claims. The original results were derived from the primary training configuration, but to address this concern we have re-executed the AE training across multiple random seeds and incorporated error bars (mean and standard deviation) into the updated NMSE, overhead, and latency figures. We have also added sensitivity analysis by evaluating the compression ratio across varied channel conditions (different room layouts and scatterer densities generated via Sionna RT). These additions, now included in the revised Results section, confirm that the 1/4 ratio consistently provides the best accuracy/feedback-size tradeoff and that the latency/throughput conclusions hold under the tested variations. revision: yes

Circularity Check

0 steps flagged

No significant circularity in derivation chain

full rationale

The paper's claims of >50% overhead reduction and throughput gains from AE CSI compression are obtained by training an autoencoder on Sionna RT-generated channels and measuring end-to-end metrics in a separate event-driven simulator against IEEE 802.11 baselines. No equations, fitted parameters, or self-citations reduce these measured outcomes to quantities defined by the same inputs; the evaluation pipeline is externally benchmarked and self-contained.

Axiom & Free-Parameter Ledger

1 free parameters · 1 axioms · 0 invented entities

The central claim rests on the assumption that simulated channels match real deployments and that the chosen AE architecture generalizes; no new physical entities are postulated.

free parameters (1)
  • compression ratio
    1/4 ratio selected after evaluating multiple options for accuracy/feedback tradeoff; not derived from first principles.
axioms (1)
  • domain assumption Sionna RT-generated channels accurately represent real-world Wi-Fi propagation for the evaluated scenarios
    Invoked to justify the simulation-based performance claims in the abstract.

pith-pipeline@v0.9.0 · 5530 in / 1175 out tokens · 27041 ms · 2026-05-10T12:34:42.386434+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

25 extracted references · 6 canonical work pages

  1. [1]

    Wi-Fi: Twenty-five years and counting,

    G. Geraci, F. Meneghello, F. Wilhelmi, D. Lopez-Perez, I. Val, L. G. Giordano, C. Cordeiro, M. Ghosh, E. Knightly, and B. Bellalta, “Wi-Fi: Twenty-five years and counting,” 2025. [Online]. Available: https://arxiv.org/abs/2507.09613

    work page arXiv 2025

  2. [2]

    IEEE 802.11be network throughput optimization with multilink operation and AP controller,

    L. Zhang, H. Yin, S. Roy, L. Cao, X. Gao, and V . Sathya, “IEEE 802.11be network throughput optimization with multilink operation and AP controller,”IEEE Internet of Things Journal, vol. 11, pp. 23 850– 23 861, 7 2024

    2024

  3. [3]

    1–502, Aug 2025

    “IEEE P802.11bn/D1.0: Draft Standard for Information technol- ogy—Telecommunications and information exchange between systems Local and metropolitan area networks—Specific requirements Part 11: Wireless LAN Medium Access Control (MAC) and Physical Layer (PHY) Specifications—Amendment 6: Enhancements for ultra-high re- liability (UHR),” pp. 1–502, Aug 2025

    2025

  4. [4]

    What will Wi-Fi 8 be? a primer on IEEE 802.11bn ultra high reliability,

    L. Galati-Giordano, G. Geraci, M. Carrascosa, and B. Bellalta, “What will Wi-Fi 8 be? a primer on IEEE 802.11bn ultra high reliability,”IEEE Communications Magazine, vol. 62, no. 8, pp. 126–132, 2024

    2024

  5. [5]

    A survey on multi-AP coordination approaches over emerging WLANs: Future directions and open challenges,

    S. Verma, T. K. Rodrigues, Y . Kawamoto, M. M. Fouda, and N. Kato, “A survey on multi-AP coordination approaches over emerging WLANs: Future directions and open challenges,”IEEE Communications Surveys & Tutorials, vol. 26, no. 2, p. 858–889, 2024. [Online]. Available: http://dx.doi.org/10.1109/COMST.2023.3344167

    work page doi:10.1109/comst.2023.3344167 2024

  6. [6]

    Zero-MUI coordinated beamforming,

    D. Ezri, S. Shilo, and R. Keren, “Zero-MUI coordinated beamforming,” November 2023, doc.: IEEE 802.11-23/1998r0

    2023

  7. [7]

    Coordinated spatial nulling (C-SN) simulations,

    R. Strobel, S. Schelstraete, I. Val, and M. Martinez, “Coordinated spatial nulling (C-SN) simulations,” January 2024, doc.: IEEE 802.11- 23/0012r0

    2024

  8. [8]

    Multi-agent reinforcement learning approach for interference optimization in Wi-Fi 8,

    J. Bacha, A. Zubow, and F. Dressler, “Multi-agent reinforcement learning approach for interference optimization in Wi-Fi 8,” 2025

    2025

  9. [9]

    Matloff, A

    N. Matloff, A. König, M. Müller, and The SimPy Development Team, SimPy: Discrete-event simulation for Python, 2023, version 4.1.1. [Online]. Available: https://simpy.readthedocs.io/

    2023

  10. [10]

    Sionna RT: Differentiable ray tracing for radio propagation modeling,

    J. Hoydis, F. A. Aoudia, S. Cammerer, M. Nimier-David, N. Binder, G. Marcus, and A. Keller, “Sionna RT: Differentiable ray tracing for radio propagation modeling,”arXiv:2303.11103, 2023

    work page arXiv 2023

  11. [11]

    Deep learning-based massive MIMO CSI feedback,

    J. Li, Q. Zhang, X. Xin, Y . Tao, Q. Tian, F. Tian, D. Chen, Y . Shen, G. Cao, Z. Gao, and J. Qian, “Deep learning-based massive MIMO CSI feedback,” in2019 18th International Conference on Optical Communications and Networks (ICOCN), 2019, pp. 1–3

    2019

  12. [12]

    Multi-resolution CSI feedback with deep learning in massive MIMO system,

    Z. Lu, J. Wang, and J. Song, “Multi-resolution CSI feedback with deep learning in massive MIMO system,” inICC 2020 - 2020 IEEE International Conference on Communications (ICC). IEEE, 6 2020

    2020

  13. [13]

    LB-SciFi: Online learning- based channel feedback for MU-MIMO in wireless LANs,

    S. Gorinsky, R. Guerin, and P. Steenkiste, “LB-SciFi: Online learning- based channel feedback for MU-MIMO in wireless LANs,” in2020 IEEE 28th International Conference on Network Protocols (ICNP). IEEE, 10 2020

    2020

  14. [14]

    Deep learning- based CSI feedback in Wi-Fi systems,

    F. Qi, J. Guo, Y . Cui, X. Li, C.-K. Wen, and S. Jin, “Deep learning- based CSI feedback in Wi-Fi systems,” in2024 5th International Symposium on Computer Engineering and Intelligent Communications (ISCEIC), 7 2024. [Online]. Available: http://arxiv.org/abs/2407.05905

    work page arXiv 2024

  15. [15]

    Impact of explicit channel sounding period on the Wi-Fi MU-MIMO performance,

    E. Endovitskiy, S. Tutelian, K. Chemrov, V . Loginov, and E. Khorov, “Impact of explicit channel sounding period on the Wi-Fi MU-MIMO performance,” in2024 IEEE International Black Sea Conference on Communications and Networking, BlackSeaCom 2024. Institute of Electrical and Electronics Engineers Inc., 2024, pp. 78–83

    2024

  16. [16]

    Cell-edge- aware precoding for downlink massive MIMO cellular networks,

    H. H. Yang, G. Geraci, T. Q. S. Quek, and J. G. Andrews, “Cell-edge- aware precoding for downlink massive MIMO cellular networks,”IEEE Transactions on Signal Processing, vol. 65, no. 13, pp. 3344–3358, 2017

    2017

  17. [17]

    1–5956, 2025

    “IEEE standard for information technology–telecommunications and information exchange between systems local and metropolitan area networks–specific requirements part 11: Wireless LAN medium access control (MAC) and physical layer (PHY) specifications,” in: IEEE Std 802.11-2024 (Revision of IEEE Std 802.11-2020), pp. 1–5956, 2025

    2024

  18. [18]

    Beamforming matrix quantization with variable feedback rate,

    C. Yuen, S. Sun, M. M. S. Ho, and Z. Zhang, “Beamforming matrix quantization with variable feedback rate,”Eurasip Journal on Wireless Communications and Networking, vol. 2012, 2012

    2012

  19. [19]

    Deep autoencoder-based massive MIMO CSI feedback with quantization and entropy coding,

    S. Ravula and S. Jain, “Deep autoencoder-based massive MIMO CSI feedback with quantization and entropy coding,” inProceedings - IEEE Global Communications Conference, GLOBECOM, 2021

    2021

  20. [20]

    Variational image compression with a scale hyperprior

    J. Ballé, D. Minnen, S. Singh, S. J. Hwang, and N. Johnston, “Variational image compression with a scale hyperprior,” inInternational Conference on Learning Representations 2018, 5 2018. [Online]. Available: http://arxiv.org/abs/1802.01436

    work page Pith review arXiv 2018

  21. [21]

    Compressai: a pytorch library and evalua- tion platform for end-to-end compression research.arXiv preprint arXiv:2011.03029, 2020

    J. Bégaint, F. Racapé, S. Feltman, and A. Pushparaja, “Compressai: a pytorch library and evaluation platform for end-to-end compression research,”arXiv preprint arXiv:2011.03029, 2020

    work page arXiv 2011

  22. [22]

    Adam: A method for stochastic optimization,

    D. Kingma and J. Ba, “Adam: A method for stochastic optimization,” International Conference on Learning Representations, 12 2014

    2014

  23. [23]

    Variable code size autoencoder (VCSA) meets CSI compression in model generalization,

    Y . Song, J. Roa, R. Zhao, Z. Rong, W. Xiao, J. Jalali, and B. Sheen, “Variable code size autoencoder (VCSA) meets CSI compression in model generalization,” in2024 International Conference on Computing, Networking and Communications, ICNC 2024. Institute of Electrical and Electronics Engineers Inc., 2 2024, pp. 209–214

    2024

  24. [24]

    Attention-infused autoencoder for massive MIMO CSI compression,

    K. Lou and X. Wu, “Attention-infused autoencoder for massive MIMO CSI compression,”IEEE Transactions on Wireless Communications, pp. 8005–8017, 11 2025. [Online]. Available: http://arxiv.org/abs/2504. 12440

    2025

  25. [25]

    Study on AI CSI compression,

    Z. Guo, P. Liu, W. Zhang, J. Yu, M. Gan, and X. Yang, “Study on AI CSI compression,” September 2023, doc.: IEEE 802.11-23/0290r4

    2023