arxiv: 2511.09464 · v2 · submitted 2025-11-12 · 📡 eess.SP

Scalable Long-Term Beamforming for Massive Multi-User MIMO

Ali Rasteh , Amirreza Kiani , Marco Mezzavilla , Sundeep Rangan This is my paper

Pith reviewed 2026-05-17 22:18 UTC · model grok-4.3

classification 📡 eess.SP

keywords massive MIMOlong-term beamformingprojection matrixcapacity upper boundmatrix inverse square rootmulti-user MIMOchannel estimationSINR bounds

0 comments

The pith

The optimal projection matrix for long-term beamforming in massive MIMO is computed using a matrix inverse square root to maximize a capacity upper bound.

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

This paper addresses scaling digital beamforming for massive MIMO systems with over 1000 antennas by focusing on long-term beamforming. It computes the projection matrix via a matrix inverse square root to maximize a capacity upper bound. The low-dimensional subspace is tracked slowly from long-term channel parameters, cutting channel estimation overhead and computation demands. Efficient matrix-matrix multiply implementations are given for hardware, with SINR error bounds derived and ray-tracing simulations confirming near-optimal performance.

Core claim

In the massive MIMO setting, we show how to optimally compute the projection matrix to maximize a capacity upper-bound using a matrix inverse square root. Computationally efficient methods are then presented to perform the matrix computation. The methods can be realized with matrix-matrix multiplies, making them amenable to systolic array implementations in hardware. Error analysis bounds on the degradation in the SINR for users are derived. Ray tracing simulations in a realistic rural uplink setting show minimal loss relative to complete instantaneous MMSE beamforming while offering significant overhead and computational gains.

What carries the argument

The projection matrix obtained via matrix inverse square root, which projects received signals onto a low-dimensional subspace tracked from long-term channel parameters to maximize the capacity upper bound.

Load-bearing premise

Long-term channel parameters alone suffice to track a low-dimensional subspace whose projection yields near-optimal capacity without frequent re-estimation of the full instantaneous channel.

What would settle it

Ray tracing or field measurements in a rural uplink scenario where the SINR or achievable rate with the proposed projection falls substantially below that of instantaneous MMSE beamforming.

Figures

Figures reproduced from arXiv: 2511.09464 by Ali Rasteh, Amirreza Kiani, Marco Mezzavilla, Sundeep Rangan.

**Figure 2.** Figure 2: CDF of the SINR for 10 UEs per sector, served by a [PITH_FULL_IMAGE:figures/full_fig_p006_2.png] view at source ↗

read the original abstract

Fully digital massive MIMO systems with large numbers (1000+) of antennas offer dramatically increased capacity gains from spatial multiplexing and beamforming. Designing digital receivers that can scale to these array dimensions presents significant challenges regarding both channel estimation overhead and digital computation. In the massive MIMO setting, long-term beamforming is widely-used since it offers significant reductions in both computation and channel estimation overhead. Long-term beamforming operates by projecting the data onto a low-dimensional subspace that can be tracked at a relatively slow time-scale from the long-term channel parameters. In this setting, we show how to optimally compute the projection matrix to maximize a capacity upper-bound using a matrix inverse square root. Computationally efficient methods are then presented to perform the matrix computation. The methods can be realized with matrix-matrix multiplies, making them amenable to systolic array implementations in hardware. Error analysis bounds on the degradation in the SINR for users are derived. Ray tracing simulations in a realistic rural uplink setting show minimal loss relative to complete instantaneous MMSE beamforming while offering significant overhead and computational gains.

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.

Desk Editor's Note private letter to a colleague

The paper gives a concrete matrix-inverse-square-root recipe for choosing long-term projections in massive MIMO, with systolic-friendly implementations and rural simulations showing small loss versus full MMSE.

read the letter

The main takeaway is that the authors describe how to pick the projection matrix for long-term beamforming by maximizing a capacity upper bound via the matrix inverse square root. They follow this with matrix-matrix multiply realizations that map to systolic arrays and derive SINR error bounds. Ray-tracing results in a rural uplink scenario indicate the approach stays close to instantaneous MMSE while cutting estimation and compute overhead for 1000-plus antenna arrays. That combination of a specific optimization step and hardware-oriented implementation is the clearest incremental piece over standard long-term subspace ideas. The simulations supply some practical evidence that the performance hit is small in the tested environment. The soft spot is the reliance on the capacity upper bound without visible details on its tightness or direct comparison to true ergodic rate maximization. If multi-user interference is strong or angular spreads shift quickly, the surrogate could drift from the actual optimum, and the single rural setting leaves open how general the minimal-loss claim is. The math on the inverse square root and the error analysis appear straightforward, but the abstract leaves the bound choice and validation steps implicit. This work targets readers focused on scalable massive MIMO hardware and implementation rather than pure theory. The methods are concrete enough and the scaling problem is real, so the paper deserves a serious referee who can check the bound tightness and ask for broader simulation cases. I would send it to review.

Referee Report

2 major / 2 minor

Summary. The manuscript proposes a method for long-term beamforming in massive multi-user MIMO systems with 1000+ antennas. It claims to optimally compute a low-dimensional projection matrix that maximizes a capacity upper-bound via the matrix inverse square root, provides efficient realizations based on matrix-matrix multiplies suitable for systolic arrays, derives SINR degradation bounds, and demonstrates via ray-tracing simulations in a rural uplink setting that performance loss relative to instantaneous MMSE beamforming is minimal while achieving substantial reductions in overhead and computation.

Significance. If the chosen capacity upper-bound is sufficiently tight or its maximizer correlates well with achievable ergodic rates under long-term covariance, the work offers a practical path to scalable beamforming that reduces channel estimation overhead and enables hardware-efficient implementations. The explicit error bounds and matrix-based realizations are concrete strengths that support the scalability claim.

major comments (2)

[§3] §3 (Optimal Projection Computation): The central claim rests on maximizing a capacity upper-bound via the matrix inverse square root to obtain the projection. However, no analysis or bound is provided showing that the maximizer of this surrogate is close to the maximizer of the true ergodic sum-rate, particularly under strong multi-user interference or varying angular spreads. This is load-bearing for the optimality and near-optimality assertions.
[§5] §5 (Simulation Results): The ray-tracing experiments report minimal loss versus instantaneous MMSE, but provide no quantitative rate or capacity values, no error-bar statistics, and no direct comparison against a subspace obtained by maximizing the true rate (rather than the upper-bound). This weakens support for the claim that long-term parameters alone suffice for near-optimal performance.

minor comments (2)

[Abstract] The abstract and introduction use the phrase 'minimal loss' without accompanying numerical deltas or operating conditions; adding these would improve clarity.
[Notation] Notation for the long-term covariance matrix and the projection matrix should be introduced once and used consistently; occasional redefinitions appear in later sections.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the constructive and detailed review of our manuscript on scalable long-term beamforming for massive multi-user MIMO. We address each major comment below with point-by-point responses and indicate planned revisions to the manuscript.

read point-by-point responses

Referee: [§3] §3 (Optimal Projection Computation): The central claim rests on maximizing a capacity upper-bound via the matrix inverse square root to obtain the projection. However, no analysis or bound is provided showing that the maximizer of this surrogate is close to the maximizer of the true ergodic sum-rate, particularly under strong multi-user interference or varying angular spreads. This is load-bearing for the optimality and near-optimality assertions.

Authors: We acknowledge that the manuscript does not include a theoretical analysis or bound quantifying the proximity of the upper-bound maximizer to the true ergodic sum-rate maximizer under varying interference or angular spread conditions. The upper-bound was selected specifically because it yields a closed-form optimal projection via the matrix inverse square root, enabling the efficient matrix-multiplication-based realizations and systolic-array compatibility that form a core contribution. We will revise §3 to add a discussion of the regimes (such as moderate-to-high SNR and covariance-based interference management) where this surrogate is expected to be reasonably tight, along with references to prior work on MIMO capacity bounds. This will clarify the scope of the optimality claim without overstating the theoretical guarantees. revision: yes
Referee: [§5] §5 (Simulation Results): The ray-tracing experiments report minimal loss versus instantaneous MMSE, but provide no quantitative rate or capacity values, no error-bar statistics, and no direct comparison against a subspace obtained by maximizing the true rate (rather than the upper-bound). This weakens support for the claim that long-term parameters alone suffice for near-optimal performance.

Authors: We agree that the simulation section would be strengthened by additional quantitative details. In the revised manuscript we will augment the figures and text in §5 with explicit ergodic rate values, capacity upper-bound numbers, and error-bar statistics computed over multiple independent ray-tracing runs. Regarding a direct comparison against the true-rate maximizer, exhaustive optimization of the ergodic sum-rate is computationally prohibitive in the full 1000+-antenna rural ray-tracing environment because it requires Monte-Carlo averaging over many small-scale fading realizations for every candidate projection. We will therefore add a supplementary experiment using a simplified analytical multi-user MIMO channel model (with known covariance matrices) that permits direct comparison of the two subspaces, thereby providing supporting evidence for the near-optimality claim while preserving the scalability focus of the work. revision: partial

Circularity Check

0 steps flagged

Derivation of optimal projection via capacity upper-bound maximization is self-contained with no circular reductions.

full rationale

The paper derives the projection matrix by directly maximizing a stated capacity upper bound using the matrix inverse square root operation on long-term covariance parameters. This constitutes a standard optimization step rather than a self-referential definition, fitted input renamed as prediction, or load-bearing self-citation chain. Error bounds and ray-tracing simulations provide external validation against instantaneous MMSE, confirming the central claim does not reduce to its inputs by construction. No equations in the abstract or described method exhibit the enumerated circular patterns.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 0 invented entities

Review performed on abstract only; the central claim therefore rests on standard massive-MIMO modeling assumptions whose precise statements are not visible.

axioms (1)

domain assumption Long-term channel statistics suffice to define a stable low-dimensional subspace for projection.
Stated in the abstract as the operating regime of long-term beamforming.

pith-pipeline@v0.9.0 · 5487 in / 1256 out tokens · 23440 ms · 2026-05-17T22:18:11.887569+00:00 · methodology

discussion (0)

Lean theorems connected to this paper

Citations machine-checked in the Pith Canon. Every link opens the source theorem in the public Lean library.

IndisputableMonolith/Cost/FunctionalEquation.lean washburn_uniqueness_aczel unclear

?

unclear
Relation between the paper passage and the cited Recognition theorem.

Lemma 1. For a given projection rank r, one matrix G_i that maximizes C_i(G_i) is G_i = [Q_i^{1/2} Q^{-1/2}]_r Q^{-1/2}
IndisputableMonolith/Foundation/AlphaCoordinateFixation.lean costAlphaLog_fourth_deriv_at_zero unclear

?

unclear
Relation between the paper passage and the cited Recognition theorem.

P(β) := sum_{k=0}^{d-1} β^k Q^k … J(β) := max_λ (λ p(λ,β)^2 - 1)^2

What do these tags mean?

matches: The paper's claim is directly supported by a theorem in the formal canon.
supports: The theorem supports part of the paper's argument, but the paper may add assumptions or extra steps.
extends: The paper goes beyond the formal theorem; the theorem is a base layer rather than the whole result.
uses: The paper appears to rely on the theorem as machinery.
contradicts: The paper's claim conflicts with a theorem or certificate in the canon.
unclear: Pith found a possible connection, but the passage is too broad, indirect, or ambiguous to say the theorem truly supports the claim.

Forward citations

Cited by 2 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score.

Low-rank Preconditioning in Beamspace Domain For Massive MU-MIMO Long-Term Beamforming
eess.SP 2026-05 unverdicted novelty 5.0

Low-rank preconditioner from top eigenpairs of the covariance matrix via randomized EVD with QRC, applied in beamspace, reduces CG iterations by 2-3x for long-term beamforming while matching exact inversion SINR.
Interference Suppression for Massive MU-MIMO Long-Term Beamforming with Matrix Inversion Approximation
eess.SP 2026-04 unverdicted novelty 5.0

Subspace nulling on long-term statistics preconditions the LTBF covariance matrix to reduce CG iterations and improve numerical stability in massive MU-MIMO.

Reference graph

Works this paper leans on

32 extracted references · 32 canonical work pages · cited by 2 Pith papers

[1]

T. L. Marzetta, E. G. Larsson, H. Yang, and H. Q. Ngo,Fundamentals of massive MIMO. Cambridge University Press, 2016

work page 2016
[2]

Massive mimo for next generation wireless systems,

E. G. Larsson, O. Edfors, F. Tufvesson, and T. L. Marzetta, “Massive mimo for next generation wireless systems,”IEEE communications magazine, vol. 52, no. 2, pp. 186–195, 2014

work page 2014
[3]

Massive MIMO evolution toward 3GPP release 18,

H. Jin, K. Liu, M. Zhang, L. Zhang, G. Lee, E. N. Farag, D. Zhu, E. Ong- gosanusi, M. Shafi, and H. Tataria, “Massive MIMO evolution toward 3GPP release 18,”IEEE Journal on Selected Areas in Communications, vol. 41, no. 6, pp. 1635–1654, 2023

work page 2023
[4]

Extreme Massive MIMO for Macro Cell Capacity Boost in 5G-Advanced and 6G,

H. V . Harri Holma and P. Mogensen, “Extreme Massive MIMO for Macro Cell Capacity Boost in 5G-Advanced and 6G,” Nokia, White Paper, 2025. [Online]. Available: https://www.nokia.com/asset/210786/

work page 2025
[5]

Energy efficient extreme MIMO: Design goals and directions,

S. Wesemann, J. Du, and H. Viswanathan, “Energy efficient extreme MIMO: Design goals and directions,”IEEE Communications Magazine, vol. 61, no. 10, pp. 132–138, 2023. 4 3 2 1 0 1 2 3 SINR (dB) 0.2 0.4 0.6 0.8CDF SINR CDF-16x16 BS Array Instantaneous Long-term BF-Exact Long-term BF-d=2 Long-term BF-d=3 (a) CDF of SINR for a16×16BS array. 3 2 1 0 1 2 3 SINR...

work page 2023
[6]

Cellular wireless networks in the upper mid-band,

S. Kang, M. Mezzavilla, S. Rangan, A. Madanayake, S. B. Venkatakrish- nan, G. Hellbourg, M. Ghosh, H. Rahmani, and A. Dhananjay, “Cellular wireless networks in the upper mid-band,”IEEE Open Journal of the Communications Society, vol. 5, pp. 2058–2075, 2024

work page 2058
[7]

Coverage evaluation of 7–15 ghz bands from existing sites,

Nokia, “Coverage evaluation of 7–15 ghz bands from existing sites,” Nokia, White Paper, 2025, accessed October 30, 2025. [Online]. Available: https://www.nokia.com/asset/213702/

work page 2025
[8]

Joint Detection, Channel Estimation and Interference Nulling for Terrestrial-Satellite Downlink Co-Existence in the Upper Mid-Band,

S. Jia, M. Ying, M. Mezzavilla, D. Calin, T. S. Rappaport, and S. Ran- gan, “Joint Detection, Channel Estimation and Interference Nulling for Terrestrial-Satellite Downlink Co-Existence in the Upper Mid-Band,” arXiv preprint arXiv:2510.08824, 2025

work page arXiv 2025
[9]

Bandwidth Gain: The Missing Gain of Massive MIMO,

M. Akrout, V . Shyianov, F. Bellili, A. Mezghani, and R. W. Heath, “Bandwidth Gain: The Missing Gain of Massive MIMO,” inICC 2023- IEEE International Conference on Communications. IEEE, 2023, pp. 5997–6003

work page 2023
[10]

A scal- able generator for massive mimo baseband processing systems with beamspace channel estimation,

Y . Dai, H. Liew, M. E. Rasekh, S. H. Mirfarshbafan, A. Gallyas- Sanhueza, J. Dunn, U. Madhow, C. Studer, and B. Nikoli ´c, “A scal- able generator for massive mimo baseband processing systems with beamspace channel estimation,” in2021 IEEE Workshop on Signal Processing Systems (SiPS). IEEE, 2021, pp. 182–187

work page 2021
[11]

Capacity of a mobile multiple- antenna communication link in Rayleigh flat fading,

T. L. Marzetta and B. M. Hochwald, “Capacity of a mobile multiple- antenna communication link in Rayleigh flat fading,”IEEE transactions on Information Theory, vol. 45, no. 1, pp. 139–157, 2002

work page 2002
[12]

Interplay of spectral efficiency, power and doppler spectrum for reference-signal-assisted wireless communication,

A. Lozano, “Interplay of spectral efficiency, power and doppler spectrum for reference-signal-assisted wireless communication,”IEEE Transac- tions on Wireless Communications, vol. 7, no. 12, pp. 5020–5029, 2008

work page 2008
[13]

Hybrid beam- forming for millimeter wave systems using the mmse criterion,

T. Lin, J. Cong, Y . Zhu, J. Zhang, and K. B. Letaief, “Hybrid beam- forming for millimeter wave systems using the mmse criterion,”IEEE Transactions on Communications, vol. 67, no. 5, pp. 3693–3708, 2019

work page 2019
[14]

Zero-forcing methods for downlink spatial multiplexing in multiuser mimo channels,

Q. H. Spencer, A. L. Swindlehurst, and M. Haardt, “Zero-forcing methods for downlink spatial multiplexing in multiuser mimo channels,” IEEE transactions on signal processing, vol. 52, no. 2, pp. 461–471, 2004

work page 2004
[15]

Long-term transmit beamforming for wireless multicast- ing,

A. Lozano, “Long-term transmit beamforming for wireless multicast- ing,” in2007 IEEE International Conference on Acoustics, Speech and Signal Processing-ICASSP’07, vol. 3. IEEE, 2007, pp. III–417

work page 2007
[16]

Space-time transmit precoding with imperfect feedback,

E. Visotsky and U. Madhow, “Space-time transmit precoding with imperfect feedback,”IEEE transactions on Information Theory, vol. 47, no. 6, pp. 2632–2639, 2002

work page 2002
[17]

Transmitter optimization and optimality of beamforming for multiple antenna systems,

S. A. Jafar and A. Goldsmith, “Transmitter optimization and optimality of beamforming for multiple antenna systems,”IEEE Transactions on Wireless Communications, vol. 3, no. 4, pp. 1165–1175, 2004

work page 2004
[18]

Physical layer multicasting in massive mimo systems with statistical csit,

K.-X. Li, L. You, J. Wang, and X. Gao, “Physical layer multicasting in massive mimo systems with statistical csit,”IEEE Transactions on Vehicular Technology, vol. 69, no. 2, pp. 1651–1665, 2019

work page 2019
[19]

An efficient global algorithm for single-group mul- ticast beamforming,

C. Lu and Y .-F. Liu, “An efficient global algorithm for single-group mul- ticast beamforming,”IEEE Transactions on Signal Processing, vol. 65, no. 14, pp. 3761–3774, 2017

work page 2017
[20]

Long-term rate-fairness-aware beamforming based massive mimo sys- tems,

W. Zhu, H. D. Tuan, E. Dutkiewicz, Y . Fang, H. V . Poor, and L. Hanzo, “Long-term rate-fairness-aware beamforming based massive mimo sys- tems,”IEEE Transactions on Communications, vol. 72, no. 4, pp. 2386– 2398, 2023

work page 2023
[21]

Beamspace channel estimation for massive mimo mmwave systems: Algorithm and vlsi design,

S. H. Mirfarshbafan, A. Gallyas-Sanhueza, R. Ghods, and C. Studer, “Beamspace channel estimation for massive mimo mmwave systems: Algorithm and vlsi design,”IEEE Transactions on Circuits and Systems I: Regular Papers, vol. 67, no. 12, pp. 5482–5495, 2020

work page 2020
[22]

Beamspace mimo for high-dimensional mul- tiuser communication at millimeter-wave frequencies,

A. Sayeed and J. Brady, “Beamspace mimo for high-dimensional mul- tiuser communication at millimeter-wave frequencies,” in2013 IEEE global communications conference (GLOBECOM). IEEE, 2013, pp. 3679–3684

work page 2013
[23]

Low- complexity polynomial channel estimation in large-scale mimo with ar- bitrary statistics,

N. Shariati, E. Bj ¨ornson, M. Bengtsson, and M. Debbah, “Low- complexity polynomial channel estimation in large-scale mimo with ar- bitrary statistics,”IEEE Journal of Selected Topics in Signal Processing, vol. 8, no. 5, pp. 815–830, 2014

work page 2014
[24]

Fast matrix inversion methods based on chebyshev and newton iterations for zero forcing precoding in massive mimo systems,

S. Hashima and O. Muta, “Fast matrix inversion methods based on chebyshev and newton iterations for zero forcing precoding in massive mimo systems,”EURASIP Journal on Wireless Communications and Networking, vol. 2020, no. 1, p. 34, 2020

work page 2020
[25]

Approximate matrix inversion for high-throughput data detection in the large-scale mimo uplink,

M. Wu, B. Yin, A. V osoughi, C. Studer, J. R. Cavallaro, and C. Dick, “Approximate matrix inversion for high-throughput data detection in the large-scale mimo uplink,” in2013 IEEE international symposium on circuits and systems (ISCAS). IEEE, 2013, pp. 2155–2158

work page 2013
[26]

Linear pre- coding based on polynomial expansion: Large-scale multi-cell mimo systems,

A. Kammoun, A. M ¨uller, E. Bj ¨ornson, and M. Debbah, “Linear pre- coding based on polynomial expansion: Large-scale multi-cell mimo systems,”IEEE Journal of Selected Topics in Signal Processing, vol. 8, no. 5, pp. 861–875, 2014

work page 2014
[27]

R. W. Heath Jr and A. Lozano,Foundations of MIMO communication. Cambridge University Press, 2018

work page 2018
[28]

Long-term channel statistic estimation for highly-mobile hybrid mmWave multi-user MIMO systems,

J. Yu, X. Liu, H. Qi, and Y . Gao, “Long-term channel statistic estimation for highly-mobile hybrid mmWave multi-user MIMO systems,”IEEE Transactions on Vehicular Technology, vol. 69, no. 12, pp. 14 277– 14 289, 2020

work page 2020
[29]

5g evolution: A view on 5g cellular technology beyond 3gpp release 15,

A. Ghosh, A. Maeder, M. Baker, and D. Chandramouli, “5g evolution: A view on 5g cellular technology beyond 3gpp release 15,”IEEE access, vol. 7, pp. 127 639–127 651, 2019

work page 2019
[30]

R. A. Horn and C. R. Johnson,Matrix analysis. Cambridge university press, 2012

work page 2012
[31]

M. J. D. Powell,Approximation theory and methods. Cambridge university press, 1981

work page 1981
[32]

Introduction to approximation theory,

E. W. Cheney, “Introduction to approximation theory,”(No Title), 1966

work page 1966