pith. sign in

arxiv: 2605.02631 · v1 · submitted 2026-05-04 · 📡 eess.SP

Multi-User XR Offloading via Massive MIMO: A System-Level Analysis using a Real-Life Dataset

Love B\'ar\'any , Ilayda Yaman , Ove Edfors , Amir Aminifar , Liang Liu This is my paper

Pith reviewed 2026-05-08 17:48 UTC · model grok-4.3

classification 📡 eess.SP
keywords XR offloadingMassive MIMOSLAMlatencylocalization errortransmission powermulti-usersystem-level analysis
0
0 comments X

The pith

Massive MIMO allows offloading of SLAM from multiple XR devices while exposing clear trade-offs among latency, localization accuracy, and transmission power.

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

The paper sets out a framework that models end-to-end latency and localization error when SLAM computations for several XR headsets are sent over a Massive MIMO wireless link to a remote server. Using parameters drawn from a real-life dataset, the authors run a case study that varies the number of users, antennas, and transmission settings. The resulting curves show how latency drops with more antennas yet localization error and device transmit power rise once the wireless resources are shared among too many users. A sympathetic reader cares because XR devices are power-limited and latency-sensitive; offloading only helps if the wireless leg does not erase the gains from remote computation.

Core claim

The authors introduce a system-level framework that combines Massive MIMO channel models with SLAM processing delays to predict latency and localization error for multiple simultaneous XR users, then apply it to a real-life dataset in a parameter sweep; the study finds that Massive MIMO yields lower latency and transmit power than conventional links for moderate user counts but requires explicit balancing against localization error and that a full device power budget remains to be evaluated.

What carries the argument

The system-level analysis framework that couples Massive MIMO uplink models with SLAM execution time and localization error to compute end-to-end performance metrics for multiple XR devices.

If this is right

  • For a moderate number of users, Massive MIMO reduces both latency and device transmission power relative to conventional cellular links while keeping localization error within acceptable bounds.
  • Increasing the number of base-station antennas improves the latency-power operating point until the per-user rate is constrained by shared resources.
  • Localization error grows once the wireless link forces compression or packet loss, creating a hard upper limit on how many users can be supported simultaneously.
  • Complete device-level power accounting, including sensing and local processing, is still required before claiming net energy savings.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The same framework could be reused to compare offloading of other XR workloads such as rendering or hand tracking.
  • Edge servers co-located with the Massive MIMO array would further shrink the propagation and queuing delays that remain after the wireless hop.
  • Antenna and power-allocation policies could be tuned in real time using the framework outputs to keep each user inside its latency and accuracy envelope.
  • If the localization-error model proves accurate, operators could dimension future XR cells by solving for the maximum user density that meets a target error threshold.

Load-bearing premise

The mathematical models of latency and localization error correctly predict real Massive MIMO link behavior and SLAM accuracy once image or feature data travel over the wireless channel.

What would settle it

An over-the-air experiment that records measured end-to-end latency, localization error, and device transmit power for several XR devices sending SLAM data through a Massive MIMO base station and compares those values directly to the framework predictions.

Figures

Figures reproduced from arXiv: 2605.02631 by Amir Aminifar, Ilayda Yaman, Liang Liu, Love B\'ar\'any, Ove Edfors.

Figure 1
Figure 1. Figure 1: Overall offloaded XR diagram. Vision SLAM components are simplified for illustrative purposes, and are only an example of the specific tasks and view at source ↗
Figure 2
Figure 2. Figure 2: Frame structures compared: A (top) and B (bottom). 1200 subcarriers view at source ↗
Figure 4
Figure 4. Figure 4: Normalised localisation error versus BER by scenario. Bootstrap view at source ↗
Figure 3
Figure 3. Figure 3: Average pose correction latency by term in (1) across all trajectories, view at source ↗
Figure 5
Figure 5. Figure 5: Average uncoded BER of Massive MIMO channel with 10 users versus view at source ↗
read the original abstract

SLAM is one of the biggest bottlenecks of XR devices, which have strict requirements for latency, power consumption, and user satisfaction. A solution that has been proposed and studied to meet the requirements is to offload SLAM to a remote server, which leverages computational hardware but may suffer due to incurred delays and transmission power. In this work, we propose offloading SLAM using Massive MIMO, which is attractive due to lower latencies, transmission power, and a more reliable link for multiple users. A framework for system-level analysis of latency and localisation error in multi-user offloaded XR with Massive MIMO has been proposed, and a case study with varying system-level parameters has been performed with it. The case study showed that there are important trade-offs between latency, localisation error, and device transmission power. We find that Massive MIMO is a promising technology for XR offloading, but that further evaluations including complete device power consumption are needed to get the full picture.

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

1 major / 1 minor

Summary. The paper proposes a framework for system-level analysis of latency and localization error in multi-user XR SLAM offloading via Massive MIMO. It performs a case study using a real-life dataset with parameter sweeps over system variables (e.g., number of base-station antennas, simultaneous users, SLAM processing delay), identifies trade-offs among latency, localization error, and device transmission power, and concludes that Massive MIMO is promising for XR offloading while noting the need for further evaluations that include complete device power consumption.

Significance. If the underlying models prove accurate, the work offers practical system-level insights into multi-user XR offloading trade-offs, leveraging a real-life dataset for the SLAM component. This could help guide wireless and computational resource allocation in emerging XR applications, though the lack of end-to-end validation limits immediate applicability.

major comments (1)
  1. [Framework and Case Study sections] The framework description and case-study parameter sweeps rely on analytic/simulation models for latency and localization error (invoked to generate all reported trade-offs) without any hardware-in-the-loop validation against real Massive MIMO links, over-the-air measurements, or SLAM accuracy under realistic multi-user channel conditions. This is load-bearing for the central claim that Massive MIMO is promising, as the quantitative results and conclusion rest on an untested abstraction.
minor comments (1)
  1. [Abstract] The abstract provides no equations, error bars, or validation details, making it difficult to assess the strength of the reported trade-offs without the full methods.

Simulated Author's Rebuttal

1 responses · 0 unresolved

We thank the referee for the constructive feedback. We address the major comment below, clarifying the scope of our simulation-based study while incorporating additional discussion of model assumptions and limitations in the revised manuscript.

read point-by-point responses

  1. Referee: [Framework and Case Study sections] The framework description and case-study parameter sweeps rely on analytic/simulation models for latency and localization error (invoked to generate all reported trade-offs) without any hardware-in-the-loop validation against real Massive MIMO links, over-the-air measurements, or SLAM accuracy under realistic multi-user channel conditions. This is load-bearing for the central claim that Massive MIMO is promising, as the quantitative results and conclusion rest on an untested abstraction.

    Authors: We agree that hardware-in-the-loop validation would further strengthen the results. Our manuscript presents a system-level simulation framework that combines established analytic models for Massive MIMO (drawn from the wireless communications literature) with a real-life SLAM dataset for the localization component. The study is explicitly positioned as a simulation-based analysis to identify trade-offs, which is standard for initial system-level evaluations of emerging technologies like XR offloading. The SLAM error model is grounded in real data, while the wireless latency and power models use well-validated abstractions (e.g., from prior Massive MIMO papers). We have added a dedicated subsection in the Discussion to explicitly state the modeling assumptions, their grounding in existing work, and the need for future over-the-air and hardware validation. This does not change the core contribution but directly addresses the concern about untested abstractions. revision: partial

Circularity Check

0 steps flagged

No significant circularity; framework uses external dataset and standard MIMO models

full rationale

The paper proposes a system-level framework for analyzing latency and localization error in multi-user XR offloading over Massive MIMO, then performs parameter sweeps in a case study using a real-life dataset solely for the SLAM component. No load-bearing step reduces by construction to a self-defined quantity, a fitted parameter renamed as a prediction, or a self-citation chain. The central trade-off results are generated from an analytic/simulation model whose equations are not shown to be equivalent to the outputs by definition, and the manuscript does not invoke uniqueness theorems or ansatzes from prior self-work to force its conclusions. This is the normal case of an independent modeling study.

Axiom & Free-Parameter Ledger

3 free parameters · 2 axioms · 0 invented entities

The analysis rests on standard wireless channel models, SLAM error propagation formulas, and latency expressions drawn from prior literature; the case study introduces multiple simulation parameters whose values are chosen or fitted to the dataset.

free parameters (3)
  • Number of base-station antennas
    Varied as a system-level parameter in the case study to observe trade-offs.
  • Number of simultaneous users
    Varied to study multi-user effects on latency and power.
  • SLAM processing delay at server
    Model parameter affecting end-to-end latency.
axioms (2)
  • domain assumption Massive MIMO channel models and beamforming gains follow standard 3GPP or similar statistical models
    Invoked when computing transmission reliability and power for multiple users.
  • domain assumption Localization error can be modeled as a function of received data rate and packet loss
    Central to the system-level latency and error framework.

pith-pipeline@v0.9.0 · 5479 in / 1475 out tokens · 24717 ms · 2026-05-08T17:48:03.042458+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

16 extracted references · 16 canonical work pages

  1. [1]

    ILLIXR: An Open Testbed to Enable Extended Reality Systems Research,

    M. Huzaifa, R. Desai, S. Grayson, X. Jiang, Y . Jing, J. Lee, F. Lu, Y . Pang, J. Ravichandran, F. Sinclair, B. Tian, H. Yuan, J. Zhang, and S. V . Adve, “ILLIXR: An Open Testbed to Enable Extended Reality Systems Research,”IEEE Micro, vol. 42, no. 4, pp. 97–106, Jul. 2022

    work page 2022

  2. [2]

    RemoteVIO: Offloading Head Tracking in an End-to-End XR System,

    Q. Jiang, Y . Pang, W. Sentosa, S. Gao, M. Huzaifa, J. Zhang, J. Perez- Ramirez, D. Das, D. Gonzalez-Aguirre, B. Godfrey, and S. Adve, “RemoteVIO: Offloading Head Tracking in an End-to-End XR System,” inProceedings of the 16th ACM Multimedia Systems Conference, ser. MMSys ’25. New York, NY , USA: Association for Computing Machinery, Mar. 2025, pp. 101–112

    work page 2025

  3. [3]

    Edge-SLAM: Edge-Assisted Visual Simultaneous Localization and Mapping,

    A. J. Ben Ali, M. Kouroshli, S. Semenova, Z. S. Hashemifar, S. Y . Ko, and K. Dantu, “Edge-SLAM: Edge-Assisted Visual Simultaneous Localization and Mapping,”ACM Trans. Embed. Comput. Syst., vol. 22, no. 1, pp. 18:1–18:31, Oct. 2022

    work page 2022

  4. [4]

    Orchestrating Joint Offloading and Scheduling for Low- Latency Edge SLAM,

    Y . Zhang, Y . Mao, H. Wang, Z. Yu, S. Guo, J. Zhang, L. Wang, and B. Guo, “Orchestrating Joint Offloading and Scheduling for Low- Latency Edge SLAM,”IEEE Transactions on Mobile Computing, vol. 24, no. 8, pp. 6901–6917, Aug. 2025

    work page 2025

  5. [5]

    DynNetSLAM: Dynamic Visual SLAM Network Offloading,

    P. Sossalla, J. Hofer, J. Rischke, C. Vielhaus, G. T. Nguyen, M. Reisslein, and F. H. P. Fitzek, “DynNetSLAM: Dynamic Visual SLAM Network Offloading,”IEEE Access, vol. 10, pp. 116 014–116 030, 2022

    work page 2022

  6. [6]

    Robust Collaborative Visual-Inertial SLAM for Mobile Augmented Reality,

    X. Pan, G. Huang, Z. Zhang, J. Li, H. Bao, and G. Zhang, “Robust Collaborative Visual-Inertial SLAM for Mobile Augmented Reality,” IEEE Transactions on Visualization and Computer Graphics, vol. 30, no. 11, pp. 7354–7363, Nov. 2024

    work page 2024

  7. [7]

    Noncooperative Cellular Wireless with Unlimited Numbers of Base Station Antennas,

    T. L. Marzetta, “Noncooperative Cellular Wireless with Unlimited Numbers of Base Station Antennas,”IEEE Transactions on Wireless Communications, vol. 9, no. 11, pp. 3590–3600, Nov. 2010

    work page 2010

  8. [8]

    Energy and Spectral Efficiency of Very Large Multiuser MIMO Systems,

    H. Q. Ngo, E. G. Larsson, and T. L. Marzetta, “Energy and Spectral Efficiency of Very Large Multiuser MIMO Systems,”IEEE Transactions on Communications, vol. 61, no. 4, pp. 1436–1449, Apr. 2013

    work page 2013

  9. [9]

    LuViRA Dataset Validation and Discussion: Comparing Vision, Radio, and Audio Sensors for Indoor Localization,

    I. Yaman, G. Tian, E. Tegler, J. Gulin, N. Challa, F. Tufvesson, O. Edfors, K. ˚Astr¨om, S. Malkowsky, and L. Liu, “LuViRA Dataset Validation and Discussion: Comparing Vision, Radio, and Audio Sensors for Indoor Localization,”IEEE Journal of Indoor and Seamless Positioning and Navigation, vol. 2, pp. 240–250, 2024

    work page 2024

  10. [10]

    The World’s First Real-Time Testbed for Massive MIMO: Design, Implementation, and Validation,

    S. Malkowsky, J. Vieira, L. Liu, P. Harris, K. Nieman, N. Kundargi, I. C. Wong, F. Tufvesson, V . ¨Owall, and O. Edfors, “The World’s First Real-Time Testbed for Massive MIMO: Design, Implementation, and Validation,”IEEE Access, vol. 5, pp. 9073–9088, 2017

    work page 2017

  11. [11]

    ORB-SLAM3: An Accurate Open-Source Library for Visual, Visual-Inertial and Multi-Map SLAM,

    C. Campos, R. Elvira, J. J. G. Rodr ´ıguez, J. M. M. Montiel, and J. D. Tard´os, “ORB-SLAM3: An Accurate Open-Source Library for Visual, Visual-Inertial and Multi-Map SLAM,”IEEE Transactions on Robotics, vol. 37, no. 6, pp. 1874–1890, Dec. 2021

    work page 2021

  12. [12]

    Implementation of Low-Latency Signal Processing and Data Shuffling for TDD Massive MIMO Sys- tems,

    S. Malkowsky, J. Vieira, K. Nieman, N. Kundargi, I. Wong, V . ¨Owall, O. Edfors, F. Tufvesson, and L. Liu, “Implementation of Low-Latency Signal Processing and Data Shuffling for TDD Massive MIMO Sys- tems,” in2016 IEEE International Workshop on Signal Processing Systems (SiPS), Oct. 2016, pp. 260–265

    work page 2016

  13. [13]

    Spectrum efficiency and processing latency trade-offs in panel-based lis,

    L. Tinnerberg, D. Iancu, O. Edfors, L. Liu, and J. V . Alegr ´ıa, “Spectrum efficiency and processing latency trade-offs in panel-based lis,” 2025. [Online]. Available: https://arxiv.org/abs/2411.19147

    work page arXiv 2025

  14. [14]

    A Tutorial on Quantitative Trajectory Evaluation for Visual(-Inertial) Odometry,

    Z. Zhang and D. Scaramuzza, “A Tutorial on Quantitative Trajectory Evaluation for Visual(-Inertial) Odometry,” in2018 IEEE/RSJ Interna- tional Conference on Intelligent Robots and Systems (IROS), Oct. 2018, pp. 7244–7251

    work page 2018

  15. [15]

    evo: Python package for the evaluation of odometry and slam

    M. Grupp, “evo: Python package for the evaluation of odometry and slam.” https://github.com/MichaelGrupp/evo, 2017

    work page 2017

  16. [16]

    Energy-Efficient Application-Specific Instruction-Set Processor for Feature Extraction in Smart Vision Systems,

    L. Ferreira, S. Malkowsky, P. Persson, S. Karlsson, K. ˚Astr¨om, and L. Liu, “Energy-Efficient Application-Specific Instruction-Set Processor for Feature Extraction in Smart Vision Systems,” in2021 55th Asilomar Conference on Signals, Systems, and Computers, Oct. 2021, pp. 324– 328

    work page 2021