REVIEW 3 major objections 4 minor 17 references
Fully Distributed Fiber-Optic Sensing Enabled by Kalman Filtering
T0 review · 3 major / 4 minor · reviewed 2026-08-05 · deepseek-v4-flash
Pith's one-line read A Kalman filter can estimate phase at the faded points where fiber-optic sensing previously went blind, recovering about 15 times more usable sensor positions.
desk verdict A practical software-only fix for fading in phi-OTDR, with a plausible demo but a load-bearing smoothness assumption that needs a localized-event test. read the letter →
The pith
A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.
The reading
What carries the argument
The load-bearing mechanism is the spatial-domain adaptive Kalman filter KFi, whose state at each fiber position is the smoothed phase and its spatial derivative, evolved along the fiber axis with the transition matrix [[1,1],[0,1]]. This is a constant-velocity spatial model: it assumes the phase profile changes smoothly with a roughly constant slope between adjacent points. With its forgetting factor set to favor predictions, KFi interpolates phase through faded positions using information from surrounding well-measured points, converting a set of discontinuous measurements into a continuous phase map from which temperature events can be located and quantified.
What would settle it
Run a φ-OTDR measurement with a heated fiber segment shorter than the effective KFi smoothing span placed precisely at a known faded point, then compare the estimated temperature-rate trace with a co-located reference thermometer; if the event appears attenuated or absent, the smoothness assumption fails. A complementary test is to feed the filter a simulated phase profile with a sharp localized step at a faded point and measure the attenuation.
Extended reading notes
Core claim
The paper claims that faded points in coherent φ-OTDR need not be discarded: their phase can be estimated by a two-stage adaptive Kalman filter. The first filter, called KFk, tracks amplitude and phase at each sampled position along the time axis using a random-walk prediction model and an online-estimated measurement noise covariance. The second filter, called KFi, models the spatial phase profile along the fiber as a smooth curve with a constant rate of change, so that phase at a faded point is predicted from its well-measured neighbors. Experiments on a fiber with a 25 m heated section show that this pipeline yields usable phase and temperature estimates at about 146 points inside the hea
Load-bearing premise
The load-bearing premise is that the true phase profile along the fiber is spatially smooth, with a roughly constant slope between neighboring points, so a faded point can be reconstructed from its well-measured neighbors; if a real event is narrower than the filter's smoothing span and sits exactly on a faded point, it will be blurred or missed.
Editorial extensions
If this is right
- Faded points no longer create blind spots, so a single fiber can provide continuous, fully distributed sensing without hardware changes.
- Temperature profiles become much denser: 146 usable points versus 10 in the demonstrated 25 m heated section, with average fidelity close to a reference thermostat.
- The software-only nature of the method means it can be applied to existing coherent φ-OTDR systems and deployed telecom fibers.
- Because the pipeline operates on the phase field before temperature conversion, the same approach should extend to strain or acoustic sensing where fading also occurs.
- Hotspot detection along critical infrastructure becomes more reliable, since an event located exactly at a previously discarded point is no longer invisible.
Reading between the lines
- The smoothness assumption is untested for sharp, impulsive events at faded points; a natural extension would be a hybrid detector that raises an alarm when the spatial filter's innovation grows large, signalling a possible localized disturbance that the smooth model may have attenuated.
- The reported 15x density gain is measured for a 25 m quasi-static heated section; the gain for narrower or faster events, or for fibers with different fading fractions, may differ and should be quantified.
- A direct simulation test feeding KFi a phase profile with a known sharp discontinuity at a faded point would separate interpolation error from measurement noise and map the filter's effective spatial resolution.
- If polarization-diverse acquisition is already available, combining it with KFi could push usable density toward 100% even under strong fading, since the filter would then have a second independent measurement stream.
Signed reviews
Editorial analysis
A structured set of objections, weighed in public.
Referee Report
Summary. The paper proposes a Kalman-filter-based pipeline for phase-sensitive OTDR that, instead of discarding polarization-faded points, estimates phase at those locations by combining noisy measurements with predictions along both the temporal and spatial axes. A temporal Sage-Husa adaptive Kalman filter (KFk) tracks amplitude and phase at each position, and a spatial filter (KFi) refines the phase profile using a constant-velocity model. Experiments on a 190 m + 185 m fiber with a 25 m heated section report that the method yields 146 usable temperature-monitoring points versus 10 for the conventional threshold-based removal, an approximately 15x improvement in spatial density, with average temperature-rate estimates close to the chamber thermostat reading. The central claim is that all measured points along the fiber can be used for sensing.
Significance. If the result holds, the paper offers a purely software-based mitigation of polarization fading in phi-OTDR, avoiding extra hardware like polarization diversity or multi-frequency sources. The experimental demonstration is a genuine, reproducible-in-principle comparison against a conventional baseline, and the idea of using spatial correlation to recover faded points is well motivated. The strength of the paper is the demonstration that a Kalman-based interpolation can recover broad, smooth temperature features at faded locations. However, the headline claim of 'fully distributed sensing' and 'all measured points can be used' is broader than what the single wide hot-zone experiment supports; the mechanism is essentially low-pass spatial interpolation, and its behavior on localized disturbances at faded points is untested.
major comments (3)
- [Event Location / Fig. 4] The central claim that faded points can be used for sensing is only validated for a 25 m heated fiber section, which is much wider than the 0.8 m spatial resolution and the effective smoothing span of KFi. With b=0.95, KFi gives high weight to predictions from neighboring points, so at a faded point the output is largely an interpolated value. A localized perturbation (e.g., a hotspot or acoustic event narrower than the smoothing span) located exactly at a faded point would be attenuated or missed. The paper needs an experiment or simulation with a narrow perturbation placed on a faded point to show that such events are recovered; otherwise the 'all points can be used for sensing' claim overreaches the evidence.
- [Phase Estimation, Eq. (2)] The description of the forgetting factor is internally inconsistent. In Eq. (2), d_k = (1-b)/(1-b^{k+1}). For b=0.05, d_0 = 1 and d_k approaches 0.95 as k grows, meaning the filter gives full weight to the innovation at k=0 and asymptotically 95% weight to measurements. The text states 'We set b = 0.05, placing high trust in predictions at k = 0 while shifting to 95% trust in measurements over time,' which is the opposite behavior. For KFi with b=0.95, d_k asymptotically approaches 0.05, consistent with 'higher weight to predictions,' but d_0 is still 1. This mismatch between the stated mechanism and the equations must be clarified; it affects the reader's ability to understand how the filter actually balances measurements and predictions.
- [Temperature Estimation / Fig. 4] The 15x density improvement is quantified from a single experimental run with an ad hoc threshold for the conventional baseline (local maxima above the average intensity of those maxima). No error bars, repeated trials, or sensitivity analysis are provided. The conventional count of 10 points could change substantially with a different threshold choice, and the proposed count of 146 depends on the filter parameters (Q, b). Without repeated measurements or a robustness analysis, the 'approximately 15 times better spatial density' claim is not yet firmly established.
minor comments (4)
- [Eq. (1)] The measurement function h(x_k) uses complex exponentials, but the state includes amplitudes A and phases psi. The Jacobian is not given, and the treatment of phase wrapping in the innovation is not described. Adding these details would improve reproducibility.
- [Eq. (3)] The process noise covariance Q for KFi is never specified or estimated. Since the smoothing behavior depends heavily on Q, its value (or estimation procedure) should be stated.
- [General] The notation for ψ^k_{Δ̂i} and Δψ^k_{Δ̂i} in the Event Location section is hard to parse; a clear definition of the spatial difference operator and its indexing would help.
- [References] Reference [8] is the source of the linear phase-temperature relation, but this relation is not derived or briefly summarized in the current paper; adding one or two equations from [8] would make the temperature estimation section self-contained.
Circularity Check
No significant circularity: Kalman phase recovery is independent of the target temperature, with only a minor non-load-bearing self-citation for the phase-temperature conversion.
full rationale
The paper's central claim—that Kalman filtering can recover usable phase at faded points and thereby increase spatial density—is not circular. The phase estimates are produced by the KFk and KFi filters (Eqs. 1–3), which combine raw I/Q measurements with a random-walk/constant-velocity state model. The target temperature is never used as an input to these filters; the forgetting factor b and process noise are tuning parameters, not fits to the thermostat reading. The 15× density improvement is a direct consequence of the filter outputting phase values at all sampled positions, and it is not equivalent by construction to the measured temperature. The temperature-phase relation is taken from the authors' prior work [8], but it is not re-derived or fitted in this paper; it is an input to the conversion step, and the central phase-recovery claim does not depend on it. The constant-velocity assumption in Eq. 3 and the untested failure mode for localized events at faded points are validity/robustness concerns, not circularity. The only minor self-citation is [8] for the phase-temperature link, and it is not load-bearing for the main contribution, so the circularity score is low.
Assumptions & free parameters
free parameters (5)
- forgetting factor b (KFk) =
0.05
- forgetting factor b (KFi) =
0.95
- conventional threshold yth =
average intensity of local maxima of |y^0_i|
- process noise covariance Q (both filters)
- initial state and covariance
assumptions (5)
- ad hoc to paper State evolution follows a random walk x_k = x_{k-1} + w_k
- ad hoc to paper Phase spatial profile follows a constant-velocity model (Eq. 3)
- domain assumption Phase evolution is linearly proportional to cumulative temperature rate K_i (from [8])
- ad hoc to paper Temperature spatial profile is rectangular (Eq. 4)
- domain assumption Kalman filter can handle the nonlinear complex measurement via standard extension (implicit)
Cite this review
Pith. "Pith review of Fully Distributed Fiber-Optic Sensing Enabled by Kalman Filtering." pith.science (2026). https://pith.science/paper/NLCTX5FV
@misc{pith2026260803667,
author = {Pith},
title = {Pith review of: Fully Distributed Fiber-Optic Sensing Enabled by Kalman Filtering},
year = {2026},
howpublished = {\url{https://pith.science/paper/NLCTX5FV}},
note = {Machine review of arXiv:2608.03667}
}
read the original abstract
Signal fading creates points along the fiber where phase cannot be extracted, so they are conventionally discarded. Instead, we propose a Kalman-based solution for {\phi}-OTDR full-fiber monitoring. Experiments demonstrate phase and temperature estimation with approximately 15 times better spatial density than if removing these points.
Figures
Reference graph
Works this paper leans on
-
[8]
Method for conversion of optical phase to temperature for coherent φ-OTDR
R. Ermakov, F . Azendorf, H. Wang, A. Sandmann, F . Da Ros, and D. Zibar, “Method for conversion of optical phase to temperature for coherent φ-OTDR”, inCLEO: Science and Innovations, 2025, SS111_4
work page 2025
-
[1]
P . Healey, “Fading in heterodyne OTDR”,Electronics letters, vol. 20, no. 1, pp. 30–32, 1984
work page 1984
-
[2]
Y . Muanenda, “Recent advances in distributed acous- tic sensing based on phase-sensitive optical time do- main reflectometry”,Journal of Sensors, vol. 2018, no. 1, p. 3 897 873, 2018
work page 2018
-
[3]
L. B. Liokumovich, N. A. Ushakov, O. I. Kotov, M. A. Bisyarin, and A. H. Hartog, “Fundamentals of optical fiber sensing schemes based on coherent optical time domain reflectometry: Signal model under static fiber conditions”,Journal of Lightwave Technology, vol. 33, no. 17, pp. 3660–3671, 2015
work page 2015
-
[4]
Distributed optical fiber sensing: Review and perspective
P . Lu et al., “Distributed optical fiber sensing: Review and perspective”,Applied physics reviews, vol. 6, no. 4, 2019
work page 2019
-
[5]
Intent and context- aware optical networks
A. Kaszubowska-Anandarajah et al., “Intent and context- aware optical networks”, in2025 25th Anniversary Inter- national Conference on Transparent Optical Networks (ICTON), IEEE, 2025, pp. 1–2
work page 2025
-
[6]
E. Ip et al., “Distributed fiber sensor network using tele- com cables as sensing media: Technology advance- ments and applications”,Journal of optical communica- tions and networking, vol. 14, no. 1, A61–A68, 2022
work page 2022
-
[7]
A. Hartog and K. Kader,Distributed fiber optic sensor system with improved linearity, US Patent 9,170,149, 2015
work page 2015
Show all 17 references
-
[9]
Quan- tum noise limited temperature-change estimation for ϕ-OTDR employing coherent detection
H. Wang, R. Ermakov, F . Da Ros, and D. Zibar, “Quan- tum noise limited temperature-change estimation for ϕ-OTDR employing coherent detection”, in2025 Euro- pean Conference on Optical Communications (ECOC), IEEE, 2025, pp. 1–4
2025
-
[10]
Noise analysis in direct detection and coherent detec- tion phase-sensitive optical time-domain reflectometry systems
P . J. Vidal-Moreno, C. Becerril, M. R. Fernández-Ruiz, H. Martins, S. Martin-Lopez, and M. Gonzalez-Herraez, “Noise analysis in direct detection and coherent detec- tion phase-sensitive optical time-domain reflectometry systems”,Optics express, vol. 31, no. 17, pp. 27 450– 27...
2023
-
[11]
On the sensitivity of dis- tributed acoustic sensing
H. Gabai and A. Eyal, “On the sensitivity of dis- tributed acoustic sensing”,Optics letters, vol. 41, no. 24, pp. 5648–5651, 2016
2016
-
[12]
Deep learning for detection of harmful events in real-world, noisy optical fiber deploy- ments
L. Sadighi, S. Karlsson, C. Natalino, L. Wosinska, M. Ruffini, and M. Furdek, “Deep learning for detection of harmful events in real-world, noisy optical fiber deploy- ments”,Journal of Lightwave Technology, vol. 43, no. 13, pp. 6092–6101, 2025
2025
-
[13]
A distributed optical fiber sensor for temperature detection in power cables
G. Yilmaz and S. E. Karlik, “A distributed optical fiber sensor for temperature detection in power cables”,Sen- sors and Actuators A: Physical, vol. 125, no. 2, pp. 148– 155, 2006
2006
-
[14]
A review of distributed fiber–optic sens- ing in the oil and gas industry
I. Ashry et al., “A review of distributed fiber–optic sens- ing in the oil and gas industry”,Journal of Lightwave Technology, vol. 40, no. 5, pp. 1407–1431, 2022
2022
-
[15]
Real time dynamic strain monitoring of optical links using the backreflec- tion of live PSK data
H. Martins, K. Shi, B. Thomsen, S. Martin-Lopez, M. Gonzalez-Herraez, and S. Savory, “Real time dynamic strain monitoring of optical links using the backreflec- tion of live PSK data”,Optics express, vol. 24, no. 19, pp. 22 303–22 318, 2016
2016
-
[16]
Characteristics and explanations of interference fading of a ϕ-OTDR with a multi-frequency source
J. Zhou, Z. Pan, Q. Y e, H. Cai, R. Qu, and Z. Fang, “Characteristics and explanations of interference fading of a ϕ-OTDR with a multi-frequency source”,Journal of lightwave technology, vol. 31, no. 17, pp. 2947–2954, 2013
2013
-
[17]
Adaptive filtering with un- known prior statistics
A. P . Sage and G. W. Husa, “Adaptive filtering with un- known prior statistics”, inJoint Automatic Control Con- ference, 1969, pp. 760–769
1969
Reviewed August 5, 2026 · model on record in the stance chip above.
Discussion (0). Continue with ORCID to comment.