Trading Datarate for Latency in Quantum Communication

Florian Seitz; Janis N\"otzel; Zuhra Amiri

REVIEW 4 major objections 5 minor 17 references

Trading Datarate for Latency in Quantum Communication

T0 review · 4 major / 5 minor · reviewed 2026-08-12 · deepseek-v4-flash

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

Pith's one-line read Quantum-optimal receivers reach latency–data-rate operating points that classical receivers cannot.

desk verdict Well-motivated latency/datarate framing undercut by a defective estimator: the central equations don't hold, so the quantum advantage curves are unsupported. read the letter →

arxiv 2411.10259 v1 pith:EKGLWYPC submitted 2024-11-15 quant-ph eess.SP

Zuhra Amiri , Florian Seitz , Janis Nötzel This is my paper · ORCID

classification quant-pheess.SP MSC 81P4594A4094A15 PACS 03.67.Hk

keywords quantumcommunicationopticalwirelesspure-lossbosonicchannelcompoundcodesestimationhomodynedetectionlatencydatarate

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

The paper aims to establish a concrete trade-off between latency and data rate for free-space optical links modeled as pure-loss bosonic channels. One corner of the trade-off is a latency-optimized scheme that sends data immediately with compound codes, which are robust to unknown channel loss but accept a lower data rate. The other corner is a data-rate-optimized scheme that spends $n_1$ channel uses on pilot signals and feedback to estimate the channel, then transmits for $n_2$ channel uses; the paper reports that a homodyne estimator followed by a compound code for the interval $[\hat{\tau}-6\delta', \hat{\tau}]$ keeps the error below $2\epsilon$ and achieves the data rate $d(\tau)$ of Eq. (24). The key claim is that some latency–data-rate operating points are reachable only when the receiver uses optimal quantum measurement strategies rather than classical ones, which matters because it puts a quantum receiver technology choice at the center of a systems-level engineering trade-off.

What carries the argument

The machinery that carries the argument is a concatenation of a channel estimator and a compound code. The estimator is a positive operator-valued measure (POVM) built from homodyne measurements on $n_1$ copies of a pilot coherent state; because homodyne outcomes on a coherent state are Gaussian, estimating the transmissivity becomes a classical hypothesis test on a Gaussian mean, and the uncertainty interval translates to $[\hat{\tau}-6\delta', \hat{\tau}]$ through $\delta' = \delta/\sqrt{2E}$. The compound code is a family of codes that works for every channel in an interval, here analyzed at the second-order coding rate of the worst channel in the interval, so the data rate is computed as $\log M^{*}(\mathcal{N}_{\hat{\tau}-6\delta'}^{\otimes n_2}, E, \epsilon)$. The second-order coding rate formula for pure-loss bosonic channels, with entropy $g(x)$ and entropy variance $v(x)$, is what turns the narrowed interval into a concrete number of bits at latency $n_1 + n_2$.

What would settle it

Directly compute or bound the second-order coding rate of the compound pure-loss bosonic channel over an interval such as $[0.001, 1]$ and compare it with the worst-case rate used in Eq. (24). A finite-blocklength search at, say, $n_2 = 2500$ that fails to reproduce the claimed rate with a $\tau$-independent code, or finds a $\tau$-independent code exceeding it, would settle whether the plotted curves and the quantum-only operating points stand.

Watch

Extended reading notes

Core claim

The central discovery is that two elementary tasks—channel estimation and compound coding—can be concatenated into a single one-round-feedback code for the pure-loss bosonic channel, and that the quantum version of the code reaches latency–data-rate points the classical version cannot. The construction is explicit: the sender transmits $n_1$ pilot pulses, the receiver performs a homodyne measurement and forms an interval estimate for the transmissivity, and then a compound code tailored to the worst-case interval $[\hat{\tau}-6\delta', \hat{\tau}]$ carries messages for $n_2$ channel uses, with total error at most $2\epsilon$. The data rate is $d(\tau)$ from Eq. (24), evaluated with the second-order coding rate of the channel with transmissivity $\hat{\tau}-6\delta'$. Numerical results with energy $E = 10^4/2$, transmissivity $\tau = 0.01$, and interval $[0.001, 1]$ show compound coding winning at short block lengths, classical estimation and feedback saturating around 1000 channel uses, and the quantum receiver's data rate still climbing beyond 2500 channel uses. The claimed consequence is that a region of the latency–data-rate plane is only accessible with optimal quantum measurement strategies.

Load-bearing premise

The whole rate calculation depends on the unproved assumption that a single code that works for every channel in an interval achieves the same second-order data rate as the worst channel in that interval does on its own; the paper's own Discussion section concedes that a proof of this is imperative.

Editorial extensions

If this is right

If the construction is correct, a link designer can pick an operating point on the trade-off curve: compound coding alone for minimum latency, pilot-and-feedback for maximum data rate at increased latency.
Some latency–data-rate points are physically reachable only with an optimal quantum receiver; classical homodyne receivers cannot reach them.
The quantum receiver's data rate keeps increasing beyond roughly 2500 channel uses, while the classical and compound schemes saturate near 1000 channel uses, so the gap widens with block length.
The concatenated scheme splits its total error budget into estimation error and decoding error, each kept below $\epsilon$ so the total stays below $2\epsilon$; this gives two independent design margins.

Reading between the lines

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

If the unproved worst-case compound-rate assumption is later replaced by a proven formula, the numerical curves will move, but the qualitative shape—compound codes for latency, estimation for rate—would likely survive.
A direct finite-blocklength simulation of $\tau$-independent codes on the interval $[0.001,1]$ would test whether the predicted inflection point near 1000 channel uses is real or an artifact of the worst-case rate bound.
The same estimator-plus-compound-code structure could be extended to channels with thermal or phase noise, replacing homodyne detection with a joint quantum measurement; the paper itself flags this as future work.

Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Desk Editor's Note

read the letter

Hi [Name],

Short version: this preprint puts two known second-order coding-rate formulas into an estimation-then-transmit scheme for OWC and plots latency vs datarate. The framing is sensible and the figures make the qualitative tradeoff easy to see. But the math under the main curves is broken, and the authors themselves flag the missing piece.

What's new: not much conceptually—the latency-datarate tradeoff is familiar from classical wireless, and the quantum advantage reduces to the known capacity gap. Still, the paper is one of the first to spell this out for a bosonic pure-loss channel with low-latency IoT in mind, so the problem statement is worth having.

Where it falls apart:

- The estimator POVM in Eq (18) integrates over a single hyperplane S_hattau, which has Lebesgue measure zero. That makes D a zero operator, so the success probability in Eq (19) is zero, and the inequality chain (20)-(23), the datarate (24), and Fig. 2 lose their foundation. You'd need to integrate over a slab of positive width and redo the confidence interval.

- Eq (16) uses n1^{-1} where Eq (15) requires n1^{-1/2}. That understates the estimation error by a factor of sqrt(n1), so the interval widths and the resulting rates are wrong even after fixing the measure-zero issue.

- The interval [tau_hat-6delta', tau_hat] is one-sided. For the true tau to be inside, you need tau_hat >= tau. A symmetric estimator will be low about half the time, so this is not a valid confidence interval. The text appears to confuse an estimate of sqrt(tau) with an estimate of tau.

- The compound-channel second-order rate is assumed to be the worst-case rate. Section VI concedes a proof is missing. That's an honest admission, but it means the compound-code curves are also unsupported.

The citation pattern is fine: the external second-order formulas are standard and correctly attributed. The authors' own previous work on compound codes is cited appropriately, not as a loop.

Who gets value: readers interested in finite-blocklength quantum communication or OWC might find the model setup suggestive, but only as a starting point. The paper needs major revision: replace the estimator with a proper positive-width region, fix the scaling, correct the confidence interval, and prove or properly cite the compound-code second-order rate. As it stands, the central claims don't follow. I'd send it to review—it's a legitimate topic and the flaws are identifiable and fixable—but I would not accept it in this form.

Best.

Referee Report

4 major / 5 minor

Summary. The manuscript studies the latency-data-rate trade-off in optical wireless communication under channel uncertainty. Two system models are proposed: a latency-optimized scheme based on compound codes that avoids pilot transmission, and a data-rate-optimized scheme that estimates the channel transmissivity via homodyne detection of pilot pulses and then uses error-free feedback to select a code for the remaining data block. The paper claims to identify operating points that are reachable only when the receiver uses optimal quantum measurement strategies, and it presents numerical latency-data-rate curves in Fig. 2. The central derivation is a concatenated estimation-and-coding argument: an estimator produces an interval of possible transmissivity values, and a compound code is then used over that interval at a second-order coding rate. The paper also discusses an IoT application in a robot factory.

Significance. If the derivation were sound, the paper would be a useful step toward finite-blocklength, latency-aware design of quantum optical links. It correctly identifies compound coding and channel-estimation-with-feedback as two natural extremes, and it draws on established second-order coding-rate results for pure-loss bosonic channels and AWGN channels. The explicit comparison between classical homodyne-based estimation and quantum-optimal receivers is a relevant direction. However, the mathematical core of the estimation step is invalid: the estimator POVM is defined by an integral over a measure-zero set, so the key lower bound in Eq. (19) cannot hold. In addition, the scaling of the confidence interval in Eq. (16) is wrong by a factor of sqrt(n1), and the compound-channel second-order coding rate is assumed rather than proved, a limitation the authors themselves concede in Section VI. These are load-bearing errors, not presentation issues, and they invalidate the quantitative claims in Eqs. (20)-(24) and Fig. 2.

major comments (4)

[Section IV-D, Eq. (18)] The estimator POVM element D_{n1}^{\hat\tau} is defined as an integral over the set S_{\hat\tau} = { q^{n1} : \sum_i q_i = n1 \sqrt{2E}\,\hat\tau }. For continuous homodyne outcomes, this is a codimension-one set of Lebesgue measure zero in R^{n1}. Therefore the integral defines the zero operator, and tr(D_{n1}^{\hat\tau} |\sqrt{\tau}E\rangle\langle\sqrt{\tau}E|^{\otimes n1}) = 0 for every state. Consequently Eq. (19) cannot hold with any positive probability, and the chain of inequalities (20)-(23), the data rate in Eq. (24), and the latency-data-rate curves in Fig. 2 have no valid estimation step. A repair would require integrating over a positive-width slab, e.g., { |\sum_i q_i - n1\sqrt{2E}\,\hat\tau| \le \Delta }, followed by a full re-derivation of Eq. (19).
[Section IV-C, Eq. (16)] Given n1 independent Gaussian samples with variance 1/2, the confidence interval in Eq. (15) has half-width \delta = (1/\sqrt{2}) \Phi^{-1}(1-\epsilon/2) / \sqrt{n1}, which scales as n1^{-1/2}. Equation (16) instead states \delta = n1^{-1} \sqrt{1/2}\,\phi^{-1}(1-\epsilon/2), which is smaller by a factor of \sqrt{n1}. This incorrect scaling understates the estimation error and therefore the width of the subsequent interval [\hat\tau - 6\delta', \hat\tau]; the numerical rates in Fig. 2 inherit this error.
[Sections IV-A, IV-E, and VI] The compound-code rates used in Eq. (24) and in the latency-optimal curve rest on the unproved assertion that 'the second-order coding rate of the compound channel is given by the worst-case second-order coding rate of the channels making up the compound channel.' No proof or supporting reference is given, and Section VI explicitly states that 'a proof of second-order coding rate for compound codes is imperative.' References [13]-[15] address capacities or sequential decoding, not finite-blocklength second-order rates for compound classical-quantum channels. Thus the reported rates are not established even when the estimation step is repaired.
[Section IV-E, Eq. (24)] Equation (5) defines the data rate d(\tau) as an average over estimation outcomes weighted by the estimator probability p_{D,\hat\tau,\tau} = tr(D_{n1}^{\hat\tau} \hat\sigma). Equation (24), however, omits this weight and integrates only p_{E,\hat\tau,\tau} over \hat\tau. Even putting aside the zero-operator issue, Eq. (24) does not follow from Eq. (5), and it overcounts estimates that occur with negligible probability.

minor comments (5)

[Eq. (3)] The expression tr(F^{n2}(\hat\tau)_{m,\hat\tau} N_\tau(\rho_{\hat\tau,m}) is missing a closing parenthesis; it should read tr(F^{n2}_{m,\hat\tau} N_\tau(\rho_{\hat\tau,m})).
[Definition 3] The POVM definition states completeness as \sum_j \Lambda_j = 1, but the estimator in Section IV-D uses a continuous outcome set; the completeness relation should be stated as an integral for continuous POVMs.
[Section IV-B] The notation 'n_p = \alpha_1, ..., \alpha_n' is confusing: the number of pilot symbols is elsewhere called n1, and the set-like notation with alpha symbols is not defined.
[Section IV-C, Eq. (16)] Equation (16) uses \phi^{-1} where Eq. (15) uses \Phi^{-1}; the inverse Gaussian CDF and the inverse Gaussian density are different functions, and the text does not clarify which is intended.
[Section IV-E] The phrase 'finally \epsilon^2 > 0' appears to be a fragment; the preceding inequality uses the fact that (1-\epsilon)^2 \ge 1 - 2\epsilon for \epsilon \in (0,1/2), but the sentence is incomplete.

Circularity Check

0 steps flagged · score 1.0 of 10

No significant circularity: the derivation is assembled from external finite-blocklength results and explicit assumptions; the main risks are soundness gaps, not circular reductions.

full rationale

The central formulas (6) and (9) are taken from external results [8,9], and the paper does not fit any parameter to a target data set and then rename it a prediction. The data-rate formula (24) inserts the external finite-blocklength bound into the concatenated-code inequality; it does not reduce to Eq. (19) by construction. The compound-rate statement in Section IV-A ('we use the reasonable assumption that the second-order coding rate of the compound channel is given by the worst-case second-order coding rate of the channels making up the compound channel') is explicitly unproved, and Section VI concedes that 'a proof of second-order coding rate for compound codes is imperative'; this is a soundness gap, not a self-citational reduction. References [13]--[15] include current authors, but the paper does not claim that those references prove the second-order compound-rate assumption; it labels the assumption as needing future proof, so the self-citation is not load-bearing in the forbidden pattern. The serious defects in the estimation step are mathematical, not circular: Eq. (18) integrates the homodyne POVM over the measure-zero hyperplane S_tauhat, which as written makes D_{n1}^{tauhat} the zero operator and Eq. (19) unsupported; and Eq. (16) uses delta ~ n1^{-1} instead of the delta ~ n1^{-1/2} required by Eq. (15). A false equation is not an input-output equivalence, so no circular step can be quoted under the hard rules. The score of 1 reflects only the same-author references and the explicit open assumption, not a circular derivation.

Assumptions & free parameters 5 free parameters · 5 assumptions · 0 invented entities

No genuinely fitted constants appear; the simulation parameters are chosen by hand. The main ledger items are the unproved compound-rate assumption and the domain assumptions about the optical channel model.

free parameters (5)

E = 10^4/2
Pilot energy (or photon-number scale) chosen for the simulation; the estimation accuracy scales with 1/sqrt(E).
epsilon = 0.5e-5
Target error probability chosen for the simulation.
tau (evaluation transmissivity) = 0.01
Channel transmissivity used to evaluate latency and data rate in the results.
a and b = 0.001 and 1
Channel parameter interval for the compound channel, chosen for simulation.
Classical AWGN SNR P
The mapping from the optical channel to the AWGN SNR is not specified in the paper, so the classical curves cannot be reproduced.

assumptions (5)

ad hoc to paper The second-order coding rate of a compound channel equals the worst-case second-order coding rate of its constituent channels.
Assumed in Section IV-A without proof; the authors flag it as open in Section VI.
domain assumption Free-space path loss follows tau = (lambda / (4 pi d))^2.
Used in Eq. (25) for the indoor factory scenario; ignores atmospheric effects and background noise.
domain assumption The indoor factory environment has no background noise and only free-space path loss.
Stated in Section IV-F to justify neglecting background noise.
domain assumption The feedback channel is error-free and has unlimited transmitter power.
Stated in Section IV-B; simplifies the analysis but is not physically justified.
standard math Known second-order coding-rate formulas for pure-loss bosonic channels (Eq. 6) and AWGN channels (Eq. 9).
Taken from refs [8] and [9]; accepted external results.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Trading Datarate for Latency in Quantum Communication." pith.science (2026). https://pith.science/paper/EKGLWYPC

@misc{pith2026241110259,
  author       = {Pith},
  title        = {Pith review of: Trading Datarate for Latency in Quantum Communication},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/EKGLWYPC}},
  note         = {Machine review of arXiv:2411.10259}
}

read the original abstract

Low latency and high data rate performance are essential in wireless communication systems. This paper explores trade-offs between latency and data rates for optical wireless communication. We introduce a latency-optimized model utilizing compound codes as one corner case and a data rate-optimized model employing channel estimation via pilot signals and feedback before data transmission. Trade-offs between the two extremes are displayed. Most importantly, we detail operating points that can only be reached when the receiver side of the link employs optimal quantum measurement strategies. Furthermore, we propose an IoT application in a robot factory as an example scenario. Our findings reveal a trade-off between latency and data rate driven by two basic algorithms: compound codes reduce latency at the cost of data rates, while channel estimation enhances data rates at the cost of latency.

Figures

Figures reproduced from arXiv: 2411.10259 by the authors.

**Figure 2.** Datarate d(τ) vs latency n(τ) of a classical and a quantum senderreceiver pair, and the quantum compound case without any feedback for comparison. The energy is E = 104/2, the transmissivity is τ = 0.01, and the lower end of the parameter interval is a = 0.001. data rate continues to increase steadily even after 2500 channel uses. VI. DISCUSSION In this section, we will discuss our findings. The findings from the s… view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

17 extracted references · 13 canonical work pages

[13]

Optimality of sequential decoding for the bosonic compound channel,

F. Seitz and J. N ¨otzel, “Optimality of sequential decoding for the bosonic compound channel,” in IEEE International Conference on Communica- tions, 2024. ICC ’24. , 2024

work page 2024
[15]

Arbitrarily varying and compound classical-quantum channels and a note on quantum zero-error capacities,

I. Bjelakovic, H. Boche, G. J. en, and N. J., “Arbitrarily varying and compound classical-quantum channels and a note on quantum zero-error capacities,” Aydinian H., Cicalese F ., Deppe C. (eds) Information Theory, Combinatorics, and Search Theory. Lecture Notes in Computer Science , vol. 7777, 2013

2013
[1]

A survey on optical technologies for iot, smart industry, and smart infrastructures,

S. Aleksic, “A survey on optical technologies for iot, smart industry, and smart infrastructures,” Journal of Sensor and Actuator networks , vol. 8, no. 3, p. 47, 2019

work page 2019
[2]

Survey on Free Space Optical Communication: A Communication Theory Perspective,

M. A. Khalighi and M. Uysal, “Survey on Free Space Optical Communication: A Communication Theory Perspective,” IEEE Communications Surveys & Tutorials , vol. 16, no. 4, pp. 2231–2258,

work page
[3]

Ultra-reliable low latency cellular networks: Use cases, challenges and approaches,

H. Chen, R. Abbas, P. Cheng, M. Shirvanimoghaddam, W. Hardjawana, W. Bao, Y . Li, and B. Vucetic, “Ultra-reliable low latency cellular networks: Use cases, challenges and approaches,” IEEE Communications Magazine, vol. 56, no. 12, pp. 119–125, 2018

work page 2018
[4]

Line-of-sight visible light communication system design and demonstration,

K. Cui, G. Chen, Z. Xu, and R. D. Roberts, “Line-of-sight visible light communication system design and demonstration,” in 2010 7th Interna- tional Symposium on Communication Systems, Networks & Digital Signal Processing (CSNDSP 2010) , 2010, pp. 621–625

work page 2010
[5]

Operating fiber networks in the quantum limit,

J. N ¨otzel and M. Rosati, “Operating fiber networks in the quantum limit,” Journal of Lightwave Technology, vol. 41, no. 22, pp. 6865–6874, 2023

work page 2023
[6]

Optical Wireless Communication for the Internet of Things: Advances, Challenges, and Opportunities,

A. S. Hamza and T. Tripp, “Optical Wireless Communication for the Internet of Things: Advances, Challenges, and Opportunities,” in Optical Wireless Communication for the Internet of Things: Advances, Challenges, and Opportunities , 2020. [Online]. Available: https://api.semanticscholar.org/CorpusID:220754408

work page 2020

Show all 17 references

[7]

Review of ultraviolet non-line-of-sight communica- tion,

R. Yuan and J. Ma, “Review of ultraviolet non-line-of-sight communica- tion,” China Communications, vol. 13, no. 6, pp. 63–75, 2016

2016
[8]

Second-order coding rates for pure-loss bosonic channels,

M. M. Wilde, J. M. Renes, and S. Guha, “Second-order coding rates for pure-loss bosonic channels,” Quantum Information Processing , vol. 15, no. 3, p. 1289–1308, Apr. 2015. [Online]. Available: http://dx.doi.org/10.1007/s11128-015-0997-x

2015 doi
[9]

Channel coding rate in the finite blocklength regime,

Y . Polyanskiy, H. V . Poor, and S. Verdu, “Channel coding rate in the finite blocklength regime,” IEEE Transactions on Information Theory , vol. 56, no. 5, pp. 2307–2359, 2010

2010
[10]

Wilde, Quantum information theory

M. Wilde, Quantum information theory . Cambridge university press, 2013

2013
[11]

Hranilovic, Wireless optical communication systems

S. Hranilovic, Wireless optical communication systems. Springer Science & Business Media, 2006

2006
[12]

Indoor optical wireless communi- cation: potential and state-of-the-art,

H. Elgala, R. Mesleh, and H. Haas, “Indoor optical wireless communi- cation: potential and state-of-the-art,” IEEE Communications Magazine , vol. 49, no. 9, pp. 56–62, 2011

2011
[14]

Compound channel capacities under energy constraints and application,

A. Cacioppo, J. N ¨otzel, and M. Rosati, “Compound channel capacities under energy constraints and application,” in 2021 IEEE International Symposium on Information Theory (ISIT) , 2021, pp. 640–645

2021
[16]

J. C. Whitaker, M. H. Kryder, J. F. Shackelford, V . K. Tripathi, G. DeSan- tis, E. D. Fabricius, M. D. Yacoub, M. Pecht, C. Sankaran, Y . Ephraim et al., The electronics handbook . Crc Press, 2018

2018
[2014]

Available: https://ieeexplore.ieee.org/document/6844864/

[Online]. Available: https://ieeexplore.ieee.org/document/6844864/

Pith tools

Reviewed August 12, 2026 · model on record in the stance chip above.

[1] [13]

Optimality of sequential decoding for the bosonic compound channel,

F. Seitz and J. N ¨otzel, “Optimality of sequential decoding for the bosonic compound channel,” in IEEE International Conference on Communica- tions, 2024. ICC ’24. , 2024

work page 2024

[2] [15]

Arbitrarily varying and compound classical-quantum channels and a note on quantum zero-error capacities,

I. Bjelakovic, H. Boche, G. J. en, and N. J., “Arbitrarily varying and compound classical-quantum channels and a note on quantum zero-error capacities,” Aydinian H., Cicalese F ., Deppe C. (eds) Information Theory, Combinatorics, and Search Theory. Lecture Notes in Computer Science , vol. 7777, 2013

2013

[3] [1]

A survey on optical technologies for iot, smart industry, and smart infrastructures,

S. Aleksic, “A survey on optical technologies for iot, smart industry, and smart infrastructures,” Journal of Sensor and Actuator networks , vol. 8, no. 3, p. 47, 2019

work page 2019

[4] [2]

Survey on Free Space Optical Communication: A Communication Theory Perspective,

M. A. Khalighi and M. Uysal, “Survey on Free Space Optical Communication: A Communication Theory Perspective,” IEEE Communications Surveys & Tutorials , vol. 16, no. 4, pp. 2231–2258,

work page

[5] [3]

Ultra-reliable low latency cellular networks: Use cases, challenges and approaches,

H. Chen, R. Abbas, P. Cheng, M. Shirvanimoghaddam, W. Hardjawana, W. Bao, Y . Li, and B. Vucetic, “Ultra-reliable low latency cellular networks: Use cases, challenges and approaches,” IEEE Communications Magazine, vol. 56, no. 12, pp. 119–125, 2018

work page 2018

[6] [4]

Line-of-sight visible light communication system design and demonstration,

K. Cui, G. Chen, Z. Xu, and R. D. Roberts, “Line-of-sight visible light communication system design and demonstration,” in 2010 7th Interna- tional Symposium on Communication Systems, Networks & Digital Signal Processing (CSNDSP 2010) , 2010, pp. 621–625

work page 2010

[7] [5]

Operating fiber networks in the quantum limit,

J. N ¨otzel and M. Rosati, “Operating fiber networks in the quantum limit,” Journal of Lightwave Technology, vol. 41, no. 22, pp. 6865–6874, 2023

work page 2023

[8] [6]

Optical Wireless Communication for the Internet of Things: Advances, Challenges, and Opportunities,

A. S. Hamza and T. Tripp, “Optical Wireless Communication for the Internet of Things: Advances, Challenges, and Opportunities,” in Optical Wireless Communication for the Internet of Things: Advances, Challenges, and Opportunities , 2020. [Online]. Available: https://api.semanticscholar.org/CorpusID:220754408

work page 2020