Pith. sign in

REVIEW 4 major objections 5 minor 1 cited by

RayLoc: Wireless Indoor Localization via Fully Differentiable Ray-tracing

T0 review · 4 major / 5 minor · reviewed 2026-08-10 · deepseek-v4-flash

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

Pith's one-line read By making ray tracing fully differentiable, RayLoc reframes wireless indoor localization as an inverse-rendering problem, recovering target positions from CSI through gradient descent on a calibrated digital scene model.

desk verdict RayLoc makes a plausible, incremental advance by applying differentiable ray tracing to localization, but the unstated per-path delay extraction undermines the central loss functions as written. read the letter →

arxiv 2501.17881 v1 pith:Y7FRG6U5 submitted 2025-01-21 eess.SP cs.AIcs.LGcs.NI

Xueqiang Han , Tianyue Zheng , Tony Xiao Han , Jun Luo This is my paper · ORCID
classification eess.SPcs.AIcs.LGcs.NI
keywords wirelessindoorlocalizationdifferentiableraytracingchannelstateinformationinverseproblemdevice-freedevice-basedscenecalibrationGaussiansmoothing
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

RayLoc's central claim is that wireless indoor localization should be posed as the inverse of a fully differentiable ray-tracing simulation, not as blind estimation of a few propagation parameters from channel state information (CSI). Given a calibrated digital model of the room—geometry, materials, and antennas—the position of a target object or device is found by gradient descent on the mismatch between simulated and measured CSI. The paper argues this matters because conventional CSI methods discard multipath and scene details as noise, while RayLoc turns them into localization signal and unifies device-free and device-based localization in one pipeline. Experiments across four indoor scenes with a single access point show average error reductions of 0.44 m to over 1.17 m compared with six baselines.

What carries the argument

The central object is a fully differentiable ray-tracing simulator built on shooting and bouncing rays, with target positions, triangle-mesh geometry, material conductivity and permittivity, antenna patterns, and transceiver locations all as trainable parameters. Because the path from scene to CSI is differentiable, automatic differentiation supplies $\partial L/\partial I$ for the localization loss, and the paper smooths that loss by convolving it with a 2D Gaussian kernel and weighting Monte Carlo samples by a bias toward low-loss positions. The CSI ratio of two antennas cancels the random phase offset in real measurements, SMAPE is used for background calibration, and MSE localizes the target.

What would settle it

Take CSI measurements with the same 20 MHz, 128-subcarrier setup in the meeting room, convert the frequency-domain CSI to a power-delay profile, and check whether distinct peaks appear at the per-path delays $\tau_i$ that the calibrated ray tracer predicts. If the roughly 50 ns delay resolution cannot resolve those paths, then the delay terms in Eqs. (16) and (19) have no measurable input from the real CSI, undermining the gradient-based localization claim.

Watch

Extended reading notes

Core claim

The paper claims that the radio channel itself can be rendered like an image: a shooting-and-bouncing-ray simulator maps a fully parameterized scene to CSI, and localization is the inverse problem of recovering the target position that generated the observed CSI. RayLoc makes every scene parameter trainable, first calibrating background geometry and material electromagnetic properties from measured CSI, then optimizing the target position while a Gaussian-kernel-smoothed loss prevents the optimizer from stalling on plateaus and local minima. With a single three-antenna access point at 5 GHz and 20 MHz bandwidth, the system localizes passive objects (device-free) and active devices (device-based) in a meeting room, laboratory, classroom, and corridor, under both line-of-sight and non-line-of-sight conditions, reducing average localization errors by 0.44 m to over 1.17 m relative to six CSI-based baselines.

Load-bearing premise

The load-bearing premise is that the per-path delays $\tau_i$ used in the background-calibration and localization losses can be extracted from real CSI, even though the 20 MHz bandwidth resolves delays only to about 50 ns (roughly 15 m of path length), and the paper does not explain how individual room reflections are separated.

Editorial extensions

If this is right

  • Localization accuracy should improve as the environment becomes more cluttered, because each reflector is treated as information; the experiments show RayLoc's advantage grows with the number of background objects.
  • One formulation covers both device-free and device-based localization, since the target position is simply one entry in the scene parameter vector.
  • The method runs on legacy WiFi hardware: a single three-antenna AP using only 20 MHz of bandwidth is sufficient, lowering deployment cost.
  • Scene calibration is the enabling step: material properties and object positions converge from measured CSI in roughly 150 iterations, and localization quality depends on this calibration.
  • Removing the Gaussian smoothing collapses device-free median error to above 1.75 m, indicating that the smoothing is what makes gradient descent tractable.

Reading between the lines

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

  • At the reported 20 MHz bandwidth the delay resolution is about 50 ns, or roughly 15 m of path length, so individual indoor multipath components cannot be separated from the measured CSI; making the per-path-delay losses in Eqs. (16) and (19) operational would require a super-resolution or parametric delay-estimation step that the paper does not describe.
  • If the inverse-RT formulation is sound, the same differentiable scene graph could be used to estimate other hidden scene properties, such as material parameters or reflector maps, not just positions.
  • The calibration dataset is the practical bottleneck; replacing the 143-195 measurement points with a one-shot depth-sensor scan, as the paper itself suggests as future work, would be the natural test of whether RayLoc scales beyond curated scenes.
  • A reader could test the method's generality by re-running the optimization with the Gaussian smoothing removed but wider bandwidth, to separate the contribution of loss smoothing from the contribution of resolvable delays.
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. The paper proposes RayLoc, a wireless indoor localization system that frames localization as the inverse of a fully differentiable ray-tracing process. The system calibrates a high-fidelity background scene model from measured CSI, then uses gradient-based optimization with a Gaussian-smoothed loss to estimate target positions. The authors claim that this approach unifies device-free and device-based localization, and they report experimental results in four indoor environments showing lower localization errors than six CSI-based baselines, with ablations supporting the smoothing components.

Significance. If the pipeline is fully specified and reproducible, RayLoc is a noteworthy contribution: it introduces a physically interpretable alternative to fingerprinting and blind parameter estimation, unifies two previously separate localization paradigms, and reports consistent improvements over strong baselines across real-world scenes. The paper also identifies a genuine optimization difficulty--sparse gradients and local minima in the ray-tracing loss landscape--and proposes a concrete mitigation. However, the current manuscript omits several load-bearing implementation details, most importantly how per-path delays are extracted from measured 20 MHz CSI, and it does not release code or data, which limits independent verification.

major comments (4)
  1. [Sec. 3.3.2, Eq. (16), and Sec. 3.4.1, Eq. (19)] The loss functions require per-path delays tau_i 'obtained from the real-world measurements', but the index set is Ntot = Nt * Nr * Ns, which counts subcarrier/antenna pairs, not individual multipath components. A single CSI matrix does not directly expose individual path delays; with the stated 20 MHz bandwidth the delay resolution is about 50 ns, i.e., roughly 15 m of path length, so individual multipath components in a room-scale environment cannot be separated by conventional IFFT or peak picking. The paper must specify the exact algorithm that maps the measured 128-subcarrier CSI to the tau_i used in Eqs. (16) and (19), or reformulate the losses using observable aggregate quantities. Without this step, the experimental results in Sec. 5.4 cannot be reproduced or fully interpreted.
  2. [Sec. 3.2.3] The claim that 'discrete decisions, such as testing the validity of paths, are replaced with soft functions' is made in a single sentence and is never defined. Occlusion and visibility changes are exactly the points where the gradient of the simulated CSI with respect to an object position is needed, so a hand-waved smoothing technique leaves the central differentiability claim unsubstantiated. Please provide the concrete soft-visibility model and its derivative, or state the approximation used and its limits.
  3. [Sec. 3.4.2, Eqs. (22)-(24)] The gradient formula in Eq. (24) does not follow directly from the smoothed loss in Eq. (23). In Eq. (23) the bias term B = exp(-alpha * L) is declared constant for gradient computation, but Eq. (24) appears to differentiate through L_{u_j} in the product g(u_j)L_{u_j}; moreover the left-hand side, derivative of L with respect to u, is not the derivative of the Monte Carlo estimate in Eq. (23) unless additional identifications are made between u and the sampled positions u_j. Since Eq. (24) is the update rule actually used for localization, a correct derivation or a precise statement of the stochastic estimator is needed.
  4. [Sec. 3.4.2 and Sec. 5.6] The paper asserts, without proof or reference, that convolving the loss with a Gaussian kernel removes plateaus and local minima while preserving the global optimum. This property is load-bearing for the method's convergence claims, and the ablation results are only indirect evidence. The authors should either prove the claim under stated assumptions, or temper the statement and clarify that the improvement is empirical.
minor comments (5)
  1. [Fig. 10(b)] The legend lists 'UbiLocate' as a baseline, but the baseline selection in Sec. 5.2 describes SpotFi, M3, and CiFi; please clarify whether this is a typo or a different baseline.
  2. [Eq. (25)] The notation 'softmax(Bc)' with scalar Bc is unusual; softmax is conventionally defined for a vector. Please specify the intended mapping (possibly a sigmoid or a vectorized softmax) and define how Bc is expanded.
  3. [Eq. (12)] The CSI matrix H is described as having size Nt * Nr * Ns, but the displayed expression is indexed only by Nt and Nr with Ns frequency entries; the nesting and orientation should be clarified.
  4. [Sec. 5.3] The text mentions 195 measurement locations for the meeting room and later says 'we collect CSI data entries from predefined 143 measurement points'; please clarify which number was used for the calibration experiment.
  5. [Throughout] There are several typographical issues, including 'T echnologies' in the author affiliation and 'V ary' in the introduction; a proofreading pass is recommended.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: RayLoc's localization output is not equal to its calibration input, and self-citations are contextual rather than load-bearing.

full rationale

RayLoc's derivation chain is not circular. The background scene parameters I\P are calibrated (Eq. 14) against measured CSI at known receiver positions using the SMAPE losses in Eqs. 16 and 18; the calibrated background model is then held fixed while the target position P is optimized by Eq. 19 and Eq. 20 against CSI measured at new, unknown positions. The target position is not an input to the calibration loss, so the localization result is not a restatement of the calibration fit. The synthetic validation in Figs. 8a-c and the real-world CSI comparison in Fig. 8d are external consistency checks, not outputs of the same optimization loop that produces the reported localization errors. The paper's self-citations (M3 [20] and MUSE-Fi [16]) are contextual related-work references and carry no load-bearing uniqueness or derivation step; the differentiable RT formulation is built on external tools and results (Mitsuba 3 [47], Sionna RT [27,28]), and the Gaussian-smoothing device in Eqs. 21-23 is an explicitly stated algorithmic construction rather than an imported result. The strongest concern in the paper, namely that the per-path delays tau_i in Eqs. 16 and 19 may not be extractable from 20 MHz CSI, is a feasibility and validation issue, not a circularity reduction: if the delay extraction is unstated or infeasible, the method would be unsupported, but the predicted target position would still not equal the calibration or delay-extraction input by construction. The limitation acknowledged in Sec. 6.2 that scene construction requires extensive CSI measurements is an engineering limitation, not a circular step. Therefore no circular step is exhibited.

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

The paper introduces no new physical entities. Its central claim rests on the domain assumption that a ray-tracing model can emulate real CSI well enough for gradient-based inversion, plus several tuned hyperparameters. The most fragile assumption is the availability of per-path delays from coarse-bandwidth CSI, which is not justified in the text.

free parameters (6)
  • loss weights gamma1 and gamma2 = reported only as a single gamma: 2 (device-free), 0.05 (device-based)
    Eq. 19 defines two weighting terms, but Section 4 reports only one gamma per setting, leaving gamma1 and gamma2 ambiguous.
  • Gaussian kernel variance sigma0 = 0.1 device-free, 1 device-based
    Controls exploration and exploitation in the smoothed loss; chosen per setting, not derived.
  • bias parameter alpha = 1 device-free, 0.2 device-based
    Weights low-loss samples in Eq. 23; tuned per setting.
  • number of sampled positions Np = 5
    Monte Carlo sample count for the smoothed loss; set by hand.
  • learning rate and momentum = 0.03 and 0.6
    Standard optimizer hyperparameters for RMSprop.
  • background material EM properties (permittivity and conductivity) per surface = fitted to calibration CSI
    The high-fidelity background model construction fits these values from measured CSI; they are not independently verified.
assumptions (6)
  • domain assumption Ray tracing with the SBR method approximates the true wireless channel accurately for the considered scenes.
    Section 3.2.1 assumes the physical model and material parameters are sufficient to generate realistic CSI.
  • domain assumption The room's floor plan and ITU-R material values provide a coarse model that, after calibration, converges to the true scene.
    Section 3.3.1 relies on this inductive bias to avoid local minima.
  • domain assumption The CSI phase offset is common across antennas and cancels in the ratio Eq. 15.
    Used to build the amplitude and delay losses for scene calibration.
  • ad hoc to paper The Gaussian-smoothed loss Eq. 23 preserves the global optimum of the true loss while removing plateaus.
    No proof is given; support is empirical only.
  • domain assumption The target height z is known and fixed during localization.
    Localization is performed on the 2D plane X by Y only.
  • domain assumption Per-path delays tau_i can be extracted from CSI measured with 20 MHz bandwidth.
    Eqs. 16 and 19 require these delays; at 20 MHz the delay resolution is about 50 ns, so this is doubtful for indoor multipath.

how reviews work

0 comments
Cite this review

Pith. "Pith review of RayLoc: Wireless Indoor Localization via Fully Differentiable Ray-tracing." pith.science (2026). https://pith.science/paper/Y7FRG6U5

@misc{pith2026250117881,
  author       = {Pith},
  title        = {Pith review of: RayLoc: Wireless Indoor Localization via Fully Differentiable Ray-tracing},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/Y7FRG6U5}},
  note         = {Machine review of arXiv:2501.17881}
}
read the original abstract

Wireless indoor localization has been a pivotal area of research over the last two decades, becoming a cornerstone for numerous sensing applications. However, conventional wireless localization methods rely on channel state information to perform blind modelling and estimation of a limited set of localization parameters. This oversimplification neglects many sensing scene details, resulting in suboptimal localization accuracy. To address this limitation, this paper presents a novel approach to wireless indoor localization by reformulating it as an inverse problem of wireless ray-tracing, inferring scene parameters that generates the measured CSI. At the core of our solution is a fully differentiable ray-tracing simulator that enables backpropagation to comprehensive parameters of the sensing scene, allowing for precise localization. To establish a robust localization context, RayLoc constructs a high-fidelity sensing scene by refining coarse-grained background model. Furthermore, RayLoc overcomes the challenges of sparse gradient and local minima by convolving the signal generation process with a Gaussian kernel. Extensive experiments showcase that RayLoc outperforms traditional localization baselines and is able to generalize to different sensing environments.

Figures

Figures reproduced from arXiv: 2501.17881 by the authors.

Figure 1
Figure 1. Unlike traditional wireless localization performing blind estimation [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. Discrepancies between pre-change and post-change power de [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 3
Figure 3. Optimization-unfriendly loss landscape of RT. [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗
Figures from the paper (12 more)
Figure 5
Figure 5. Figure 5: Illustration of the SBR process. Each ray emitted from the Tx is defined by its origin Q0 and an initial normalized orientation vector d0. The emitted ray marches a distance t0 along its initial direction d0, interacting with the triangular primitive of an object in th…
Figure 4
Figure 4. Figure 4: The workflow of RayLoc. where 0 ≤ m1, m2 ≤ 1 are the weight for vertices. When ti−1 > 0, m1 > 0, m2 > 0 is satisfied, the intersection point Qi is within the triangle surface and can be obtained with the Eq. 1 and 2. The direction of the outgoing ray di is determined b…
Figure 5
Figure 5. Figure 5: Then, the EM field is distorted by the interactions [PITH_FULL_IMAGE:figures/full_fig_p005_5.png]
Figure 6
Figure 6. Figure 6: The enhanced loss landscape with a different number of sampled [PITH_FULL_IMAGE:figures/full_fig_p008_6.png]
Figure 8
Figure 8. Figure 8: Learning curves for the scene parameters and the CSI before [PITH_FULL_IMAGE:figures/full_fig_p009_8.png]
Figure 7
Figure 7. Figure 7: The layout of four selected experimental scenes. [PITH_FULL_IMAGE:figures/full_fig_p009_7.png]
Figure 9
Figure 9. Figure 9: Localization errors for device-free and device-based settings in [PITH_FULL_IMAGE:figures/full_fig_p010_9.png]
Figure 10
Figure 10. Figure 10: Performance under different background clutter levels. [PITH_FULL_IMAGE:figures/full_fig_p010_10.png]
Figure 12
Figure 12. Figure 12: Impact of AP height. with a step size of one. The localization errors of RayLoc for device-free and device-based configurations are reported in Fig. 13a and Fig. 13b, respectively. As shown in [PITH_FULL_IMAGE:figures/full_fig_p011_12.png]
Figure 13
Figure 13. Figure 13: Impact of AP number. 5.5.4 Impact of Variance σ The role of the variance σ in RayLoc is to adjust the relative importance of the sampling positions, determining the proportion that each sample contributes to the final gradient calculation. A larger value σ causes the …
Figure 14
Figure 14. Figure 14: Impact of variance σ. by the variations of the loss function around the current position. The variation is governed not only by the variance σ but also by the bias α. A high value of α makes the weighted gradient concentrated in a few sample positions leading to lower…
Figure 15
Figure 15. Figure 15: Impact of bias α. 5.6 Ablation Study In this section, we conduct the ablation study to evaluate the impact of each module of RayLoc on the localization performance. We compare the original RayLoc effectiveness with the following cases: removing the adaptive variance, …

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. RadioDiff-Loc: Diffusion Model Enhanced Scattering Congnition for NLoS Localization with Sparse Radio Map Estimation

    eess.SY 2025-09 conditional novelty 5.0 of 10

    A diffusion model reconstructs radio maps from sparse vertex-level RSS samples, yielding non-line-of-sight emitter localization without known transmit power.

Reference graph

Works this paper leans on

51 extracted references · 49 canonical work pages · cited by 1 Pith paper

  1. [40]

    Fully Differentiable Ray Tracing via Discontinuity Smoothing for Radio Network Opti- mization,

    J. Eertmans, L. Jacques, and C. Oestges, “Fully Differentiable Ray Tracing via Discontinuity Smoothing for Radio Network Opti- mization,” in Proc. of the 18th EuCAP , pp. 1–5

    work page

  2. [42]

    Calibrating Wireless Ray Tracing for Digital Twinning Using Local Phase Error Estimates,

    C. Ruah, O. Simeone, J. Hoydis, and B. Al-Hashimi, “Calibrating Wireless Ray Tracing for Digital Twinning Using Local Phase Error Estimates,” IEEE T ransactions on Machine Learning in Communica- tions and Networking , vol. 2, pp. 1193–1215, 2024

    work page 2024

  3. [1]

    Location-based Services: Ongoing Evolution and Re- search Agenda,

    H. Huang, G. Gartner, J. M. Krisp, M. Raubal, and N. Van de Weghe, “Location-based Services: Ongoing Evolution and Re- search Agenda,” Journal of Location Based Services , vol. 12, no. 2, pp. 63–93, 2018

    work page 2018

  4. [2]

    Geolocation API Overview,

    Google, “Geolocation API Overview,” https://developers.google. com/maps/documentation/geolocation/overview?hl=zh-cn, 2021

    work page 2021

  5. [3]

    Cisco Spaces Data Sheet,

    Cisco, “Cisco Spaces Data Sheet,” https://www.cisco. com/c/en/us/products/collateral/wireless/dna-spaces/ datasheet-c78-741786.html, 2022

    work page 2022

  6. [4]

    iBeacon,

    T. Ehrens, “iBeacon,” https://www.techtarget.com/iotagenda/ definition/Apple-iBeacon, 2019

    work page 2019

  7. [5]

    A Survey of Indoor Local- ization Systems and Technologies,

    F. Zafari, A. Gkelias, and K. K. Leung, “A Survey of Indoor Local- ization Systems and Technologies,” IEEE Communications Surveys & T utorials, vol. 21, no. 3, pp. 2568–2599, 2019

    work page 2019

  8. [6]

    Real- time Indoor Localization in Smart Homes using Semi-supervised Learning,

    N. Ghourchian, M. Allegue-Martinez, and D. Precup, “Real- time Indoor Localization in Smart Homes using Semi-supervised Learning,” in Proc. of AAAI , vol. 31, no. 2, 2017, pp. 4670–4677

    work page 2017

Show all 51 references
  1. [7]

    Deep Learning based Wireless Localization for Indoor Navigation,

    R. Ayyalasomayajula, A. Arun, C. Wu, S. Sharma, A. R. Sethi, D. Vasisht, and D. Bharadia, “Deep Learning based Wireless Localization for Indoor Navigation,” in Proc. of the 26th ACM MobiCom, 2020, pp. 1–14

    2020

  2. [8]

    Real-time Crowd Monitoring using Seamless Indoor-outdoor Localization,

    T. Kulshrestha, D. Saxena, R. Niyogi, and J. Cao, “Real-time Crowd Monitoring using Seamless Indoor-outdoor Localization,” IEEE T ransactions on Mobile Computing, vol. 19, no. 3, pp. 664–679, 2019

    2019

  3. [9]

    RADAR: An In-building RF-based User Location and Tracking System,

    P . Bahl and V . Padmanabhan, “RADAR: An In-building RF-based User Location and Tracking System,” in Proc. of the 19th IEEE INFOCOM, vol. 2, 2000, pp. 775–784 vol.2

    2000

  4. [10]

    Indoor Localization Without The Pain,

    K. Chintalapudi, A. Padmanabha Iyer, and V . N. Padmanabhan, “Indoor Localization Without The Pain,” in Proc. of the 16th ACM MobiCom, 2010, pp. 173–184

    2010

  5. [11]

    RASS: A Real-Time, Accu- rate, and Scalable System for Tracking Transceiver-Free Objects,

    D. Zhang, Y. Liu, X. Guo, and L. M. Ni, “RASS: A Real-Time, Accu- rate, and Scalable System for Tracking Transceiver-Free Objects,” IEEE T ransactions on Parallel and Distributed Systems , vol. 24, no. 5, pp. 996–1008, 2013

    2013

  6. [12]

    Radio- Frequency Tomography for Passive Indoor Multitarget Tracking,

    S. Nannuru, Y. Li, Y. Zeng, M. Coates, and B. Yang, “Radio- Frequency Tomography for Passive Indoor Multitarget Tracking,” IEEE T ransactions on Mobile Computing , vol. 12, no. 12, pp. 2322– 2333, 2013

    2013

  7. [13]

    FILA: Fine-grained Indoor Localization,

    K. Wu, J. Xiao, Y. Yi, M. Gao, and L. M. Ni, “FILA: Fine-grained Indoor Localization,” in Proc. of the 31st IEEE INFOCOM , 2012, pp. 2210–2218

    2012

  8. [14]

    SpotFi: Decimeter Level Localization using WiFi,

    M. Kotaru, K. Joshi, D. Bharadia, and S. Katti, “SpotFi: Decimeter Level Localization using WiFi,” in Proc. of ACM SIGCOMM , 2015, pp. 269–282

    2015

  9. [15]

    ArrayTrack: A Fine-Grained Indoor Location System,

    J. Xiong and K. Jamieson, “ArrayTrack: A Fine-Grained Indoor Location System,” in Proc. of the 10th USENIX NSDI . Lombard, IL: USENIX Association, Apr. 2013, pp. 71–84

    2013

  10. [16]

    MUSE-Fi: Con- tactless Muti-person Sensing Exploiting Near-field Wi-Fi Channel Variation,

    J. Hu, T. Zheng, Z. Chen, H. Wang, and J. Luo, “MUSE-Fi: Con- tactless Muti-person Sensing Exploiting Near-field Wi-Fi Channel Variation,” in Proc. of the 29th ACM MobiCom , 2023, pp. 1–15

    2023

  11. [17]

    BFMSense: WiFi Sensing Using Beamforming Feedback Matrix,

    E. Yi, D. Wu, J. Xiong, F. Zhang, K. Niu, W. Li, and D. Zhang, “BFMSense: WiFi Sensing Using Beamforming Feedback Matrix,” in Proc. of the 21st USENIX NSDI . Santa Clara, CA: USENIX Association, 2024, pp. 1697–1712

    2024

  12. [18]

    Accurate Indoor Localiza- tion With Zero Start-up Cost,

    S. Kumar, S. Gil, D. Katabi, and D. Rus, “Accurate Indoor Localiza- tion With Zero Start-up Cost,” in Proc. of the 20th ACM MobiCom , 2014, pp. 483–494

    2014

  13. [19]

    mD-Track: Leveraging Multi-dimensionality for Passive Indoor Wi-Fi Tracking,

    Y. Xie, J. Xiong, M. Li, and K. Jamieson, “mD-Track: Leveraging Multi-dimensionality for Passive Indoor Wi-Fi Tracking,” in Proc. of the 25th ACM MobiCom

  14. [20]

    M 3: Multipath Assisted Wi-Fi Localization With a Single Access Point,

    Z. Chen, G. Zhu, S. Wang, Y. Xu, J. Xiong, J. Zhao, J. Luo, and X. Wang, “ M 3: Multipath Assisted Wi-Fi Localization With a Single Access Point,” IEEE T ransactions on Mobile Computing , vol. 20, no. 2, pp. 588–602, 2021

    2021

  15. [21]

    SIABR: A Structured Intra-Attention Bidirectional Recurrent Deep Learning Method for Ultra-Accurate Terahertz Indoor Localization,

    S. Fan, Y. Wu, C. Han, and X. Wang, “SIABR: A Structured Intra-Attention Bidirectional Recurrent Deep Learning Method for Ultra-Accurate Terahertz Indoor Localization,” IEEE Journal on Selected Areas in Communications, vol. 39, no. 7, pp. 2226–2240, 2021

    2021

  16. [22]

    WILL: Wireless Indoor Localization Without Site Survey,

    C. Wu, Z. Yang, Y. Liu, and W. Xi, “WILL: Wireless Indoor Localization Without Site Survey,”IEEE T ransactions on Parallel and Distributed Systems, vol. 24, no. 4, pp. 839–848, 2013

    2013

  17. [23]

    NNE-Tracking: A Neural Network Enhanced Framework for Device-Free Wi-Fi Tracking,

    X. Tong, W. Ge, Y. Tian, Z. Liu, X. Liu, and W. Qu, “NNE-Tracking: A Neural Network Enhanced Framework for Device-Free Wi-Fi Tracking,” IEEE T ransactions on Mobile Computing , vol. 23, no. 9, pp. 8981–8998, 2024

    2024

  18. [24]

    DAFI: WiFi-based Device-free Indoor Localization via Domain Adaptation,

    H. Li, X. Chen, J. Wang, D. Wu, and X. Liu, “DAFI: WiFi-based Device-free Indoor Localization via Domain Adaptation,” Proc. of the ACM UbiComp , vol. 5, no. 4, Dec. 2022

    2022

  19. [25]

    WiNeRT: Towards Neural Ray Tracing for Wireless Chan- nel Modelling and Differentiable Simulations,

    T. Orekondy, P . Kumar, S. Kadambi, H. Ye, J. Soriaga, and A. Be- hboodi, “WiNeRT: Towards Neural Ray Tracing for Wireless Chan- nel Modelling and Differentiable Simulations,” in Proc. of the 17th ICLR, 2023

    2023

  20. [26]

    Nerf2: Neural radio- frequency radiance fields,

    X. Zhao, Z. An, Q. Pan, and L. Yang, “Nerf2: Neural radio- frequency radiance fields,” in Proceedings of the 29th Annual Inter- national Conference on Mobile Computing and Networking , 2023, pp. 1–15

    2023

  21. [27]

    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,” 2023

    2023

  22. [28]

    Learning Radio Environments by Differentiable Ray Tracing,

    J. Hoydis, F. A. Aoudia, S. Cammerer, F. Euchner, M. Nimier- David, S. T. Brink, and A. Keller, “Learning Radio Environments by Differentiable Ray Tracing,”IEEE T ransactions on Machine Learn- ing in Communications and Networking , vol. 2, pp. 1527–1539, 2024

    2024

  23. [29]

    Multiverse at the Edge: Interacting Real World and Digital Twins for Wireless Beamforming,

    B. Salehi, U. Demir, D. Roy, S. Pradhan, J. Dy, S. Ioannidis, and K. Chowdhury, “Multiverse at the Edge: Interacting Real World and Digital Twins for Wireless Beamforming,” IEEE/ACM T ransac- tions on Networking , vol. 32, no. 4, pp. 3092–3110, 2024

    2024

  24. [30]

    A DL-Based High- Precision Positioning Method in Challenging Urban Scenarios for B5G CCUAVs,

    K. Gao, H. Wang, H. Lv, and P . Gao, “A DL-Based High- Precision Positioning Method in Challenging Urban Scenarios for B5G CCUAVs,” IEEE Journal on Selected Areas in Communications , vol. 41, no. 6, pp. 1670–1687, 2023

    2023

  25. [31]

    Widar2.0: Passive Human Tracking with A Single Wi-Fi Link,

    K. Qian, C. Wu, Y. Zhang, G. Zhang, Z. Yang, and Y. Liu, “Widar2.0: Passive Human Tracking with A Single Wi-Fi Link,” in Proc. of the 16th ACM MobiSys , 2018, pp. 350–361

    2018

  26. [32]

    Effects of Building Materials and Structures on Ra- diowave Propagation Above About 100 MHz,

    P . Series, “Effects of Building Materials and Structures on Ra- diowave Propagation Above About 100 MHz,” Recommendation ITU-R, pp. 2040–1, 2015

    2015

  27. [33]

    Dielectric Material Characterization of Concrete in GHz Range in Dependence on Pore Volume and Water Content,

    T. T. Dinh, S. Hegler, M. Liebscher, I. N. de Sosa, H. Li, D. Plette- meier, W.-G. Drossel, and V . Mechtcherine, “Dielectric Material Characterization of Concrete in GHz Range in Dependence on Pore Volume and Water Content,” Construction and Building Ma- terials, vol. 311, p....

    2021

  28. [34]

    Tse and P

    D. Tse and P . Viswanath, Fundamentals of Wireless Communication . Cambridge University Press, 2005

    2005

  29. [35]

    Geo- metric Modeling Based on Triangle Meshes,

    M. Botsch, M. Pauly, C. Rossl, S. Bischoff, and L. Kobbelt, “Geo- metric Modeling Based on Triangle Meshes,” in ACM SIGGRAPH 2006 Courses, 2006, pp. 1–es

    2006

  30. [36]

    Advancing Accuracy of Shooting and Bouncing Rays Method for Ray-Tracing Propagation Modeling Based on Novel Approaches to Ray Cone Angle Calculation,

    S. Kasdorf, B. Troksa, C. Key, J. Harmon, and B. M. Notaro ˇs, “Advancing Accuracy of Shooting and Bouncing Rays Method for Ray-Tracing Propagation Modeling Based on Novel Approaches to Ray Cone Angle Calculation,” IEEE T ransactions on Antennas and Propagation, vol. 69, no. 8...

    2021

  31. [37]

    Ray Techniques in Electromagnetics,

    G. Deschamps, “Ray Techniques in Electromagnetics,” Proc. of the IEEE, vol. 60, no. 9, pp. 1022–1035, 1972

    1972

  32. [38]

    Learnable Wireless Digital Twins: Reconstructing Elec- tromagnetic Field With Neural Representations,

    S. Jiang, Q. Qu, X. Pan, A. Agrawal, R. Newcombe, and A. Alkha- teeb, “Learnable Wireless Digital Twins: Reconstructing Elec- tromagnetic Field With Neural Representations,” arXiv preprint arXiv:2409.02564, 2024

    2024 arXiv

  33. [39]

    Pix2Vex: Image-to-geometry Reconstruction Using a Smooth Dif- ferentiable Renderer,

    F. Petersen, A. H. Bermano, O. Deussen, and D. Cohen-Or, “Pix2Vex: Image-to-geometry Reconstruction Using a Smooth Dif- ferentiable Renderer,” arXiv preprint arXiv:1903.11149 , 2019

    1903 arXiv

  34. [41]

    Material and Lighting Reconstruction for Com- plex Indoor Scenes With Texture-space Differentiable Rendering,

    M. Nimier-David, Z. Dong, W. Jakob, A. Kaplanyan, A. Bousseau, and M. McGuire, “Material and Lighting Reconstruction for Com- plex Indoor Scenes With Texture-space Differentiable Rendering,” in Proc. of EGSR , 2021, pp. 73–84

    2021

  35. [43]

    FarSense: Pushing the Range Limit of WiFi-based Respiration Sensing With CSI Ratio of Two Antennas,

    Y. Zeng, D. Wu, J. Xiong, E. Yi, R. Gao, and D. Zhang, “FarSense: Pushing the Range Limit of WiFi-based Respiration Sensing With CSI Ratio of Two Antennas,” Proc. of ACM UbiComp , vol. 3, no. 3, pp. 1–26, 2019

    2019

  36. [44]

    A Simple and Efficient Smoothing Method for Faster Optimization and Local Exploration,

    K. Scaman, L. DOS SANTOS, M. Barlier, and I. Colin, “A Simple and Efficient Smoothing Method for Faster Optimization and Local Exploration,” in Proc. of NeurIPS , H. Larochelle, M. Ran- zato, R. Hadsell, M. Balcan, and H. Lin, Eds., vol. 33. Curran Associates, Inc., 2020, pp. ...

    2020

  37. [45]

    Blender 3.6,

    “Blender 3.6,” https://www.blender.org/download/releases/ 3-6/, 2023

    2023

  38. [46]

    Mitsuba Blender Add-on,

    “Mitsuba Blender Add-on,” https://github.com/ mitsuba-renderer/mitsuba-blender, 2023

    2023

  39. [47]

    Mitsuba 3 Renderer,

    W. Jakob, S. Speierer, N. Roussel, M. Nimier-David, D. Vicini, T. Zeltner, B. Nicolet, M. Crespo, V . Leroy, and Z. Zhang, “Mitsuba 3 Renderer,” 2022, https://mitsuba-renderer.org

    2022

  40. [48]

    AX-CSI: En- abling CSI Extraction on Commercial 802.11 Ax Wi-Fi Platforms,

    F. Gringoli, M. Cominelli, A. Blanco, and J. Widmer, “AX-CSI: En- abling CSI Extraction on Commercial 802.11 Ax Wi-Fi Platforms,” in Proc. of the 15th ACM Workshop on WiNTECH , 2022, pp. 46–53

    2022

  41. [49]

    WiTraj: Robust Indoor Motion Tracking With WiFi Signals,

    D. Wu, Y. Zeng, R. Gao, S. Li, Y. Li, R. C. Shah, H. Lu, and D. Zhang, “WiTraj: Robust Indoor Motion Tracking With WiFi Signals,” IEEE T ransactions on Mobile Computing, vol. 22, no. 5, pp. 3062–3078, 2023

    2023

  42. [50]

    DSCP: Depthwise Separable Convolution-Based Passive Indoor Localization Using CSI Fingerprint,

    C. Han, W. Xun, L. Sun, Z. Lin, and J. Guo, “DSCP: Depthwise Separable Convolution-Based Passive Indoor Localization Using CSI Fingerprint,” Wireless Communications and Mobile Computing , vol. 2021, no. 1, p. 8821129, 2021

    2021

  43. [51]

    CiFi: Deep Convolutional Neural Networks for Indoor Localization With 5 GHz Wi-Fi,

    X. Wang, X. Wang, and S. Mao, “CiFi: Deep Convolutional Neural Networks for Indoor Localization With 5 GHz Wi-Fi,” in Proc. of IEEE ICC, 2017, pp. 1–6

    2017

Pith tools

Reviewed August 10, 2026 · model on record in the stance chip above.