Pith. sign in

REVIEW 2 major objections 4 minor 24 references

Covert Millimeter-Wave Communication via a Dual-Beam Transmitter

T0 review · 2 major / 4 minor · reviewed 2026-08-14 · deepseek-v4-flash

Pith's one-line read A dual-beam mmWave transmitter with random jamming power can hide a positive-rate data beam from Willie, and the paper gives closed forms for the detection and outage trade-off.

desk verdict A workmanlike extension of jammer-assisted covert communication to mmWave with solid closed-form analysis, but the positive-rate claim only works if Willie is trying to detect a data beam on top of an always-on jammer, not the existence of any transmission. read the letter →

arxiv 1908.07591 v1 pith:RO6ZOZZK submitted 2019-08-20 cs.IT math.IT

classification cs.ITmath.IT MSC 94A1594A40
keywords covertcommunicationlowprobabilityofdetectionmillimeter-wavedual-beamtransmitterjammingoutageNakagamifadingwarden
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 sets out to show that covert communication—hiding the very existence of a transmission from a warden—can sustain positive rates over millimeter-wave links, in contrast to the zero-rate square-root law of ordinary AWGN covert channels. The trick is a dual-beam transmitter: one antenna array beams data to Bob, while a second array points a jamming beam at Willie with random power that changes each block. With that jammer active, Willie's optimal energy detector cannot separate 'data present' from 'data absent,' and the paper derives closed forms for Willie's expected detection error and for Bob's outage probability. The upshot is a concrete formula for the maximum covert data rate as a function of jamming power, target rate, beamwidths, and fading parameters.

What carries the argument

The load-bearing object is the dual-beam transmitter and, specifically, the independent jamming beam. Willie's optimal detector is a threshold on average received power; under $H_0$ (no data beam) Willie still sees the jammer's received power plus noise, and under $H_1$ the data term is added. Because $P_J$ is uniform and unknown, and the Alice\u2013Willie link is averaged over LOS/NLOS blockage, Nakagami fading, and array misalignment, the two hypotheses' received-power distributions overlap. The analysis chain uses the sectored antenna model, a gamma CDF approximation, and gamma moment generating functions to convert the overlap into closed-form error and outage expressions.

What would settle it

Run Willie's detector with the jammer turned off under $H_0$ (i.e., $P_J=0$ when Alice sends nothing to Bob): then the received power under $H_0$ is just the noise, while under $H_1$ it contains both jamming and data terms, so an optimal threshold can drive $P_{e,w}^*$ to $0$ for large $n$\u2014exactly the regime the paper's positive-rate claim avoids. Detecting this behavior would refute any reading of the result as hiding all transmissions from Alice.

Watch

Extended reading notes

Core claim

The central claim is that by randomizing the jamming power $P_J$ over $[0,P_J^{\max}]$ and steering it at Willie, Alice can force Willie's minimum detection error $P_{e,w}^*$ toward $1$ (i.e., $E[P_{e,w}^*]\ge 1-\epsilon$ for any $\epsilon>0$) while still sending Bob data at positive rate $R_b(1-P_{\mathrm{out}}^{AB})$. The paper proves this in the large-block regime $n\to\infty$ under optimal threshold detection: Theorem 1 gives Willie's error for fixed channels, Theorem 2 gives its expectation over blockage, beam misalignment, and fading, and Theorem 3 gives Bob's outage probability. Proposition 4 then defines the optimal covert rate by choosing the largest $P_J^{\max}$ that keeps Willie's expected error at least $1-\epsilon$. Numerical results show effective rates up to roughly $4.35$ bits per channel use for $\epsilon=0.05$ in the benchmark setting, with the caveat that mmWave's wide bandwidths translate these per-use rates into large bit-per-second gains.

Load-bearing premise

Covertness is only claimed against a warden who already knows Alice is radiating a jamming beam; under the null hypothesis the jammer stays on, so the analysis hides the data beam, not the transmitter's activity.

Editorial extensions

If this is right

  • For any $\epsilon>0$, there are parameter choices (large enough $P_J^{\max}$ and suitable beam gains) under which Alice and Bob communicate at positive rate while Willie's expected detection error stays above $1-\epsilon$.
  • The optimal jamming power for a fixed covertness requirement is the root of $E[P_{e,w}^*]=1-\epsilon$, giving $R_{a,b}^*=R_b(1-P_{\mathrm{out}}^{AB})$.
  • Increasing the jammer's main-lobe gain toward Willie or reducing Alice's side-lobe leakage toward Willie improves covertness, whereas increasing the data power $P_a$ or its side-lobe gain worsens it.
  • Because Bob receives the data beam through its main lobe and the jammer through a side lobe, mmWave beamforming separates the two roles; the same physical setup in omnidirectional RF would not achieve the same separation.
  • The derived expressions reduce to Rayleigh fading by setting the Nakagami parameter $\nu_B=1$.

Reading between the lines

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

  • The covertness guarantee is relative: Willie is assumed to know and expect the jamming beam. If Willie instead tests whether Alice is radiating at all, the jamming beam itself is a detectable transmission, so the result is better read as 'covert data on top of a public/cover jammer' rather than 'invisible transmitter.'
  • The same structure could be applied to other random jamming distributions; the paper's uniform-$P_J$ choice is convenient but not essential, and skewed distributions might improve the rate\u2013outage tradeoff.
  • At finite blocklength, fluctuations in the empirical received power will add a further penalty; the $n\to\infty$ limit here is optimistic, and quantifying that penalty is a natural next step.
  • One testable extension is to let Willie's location be random: the closed forms already average over blockage and beam misalignment, so adding warden position would directly give a spatially averaged covert rate for a network.
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.

Referee Report

2 major / 4 minor

Summary. The paper studies covert communication over millimeter-wave bands using a dual-beam transmitter: one array sends a data beam to Bob, and a second array sends an independent, block-varying jamming beam toward Willie, with jamming power uniformly distributed in [0, Pmax_J]. The authors characterize Willie's optimal-detection error probability Pe,w in closed form (Theorem 1), its expectation from Alice's perspective (Theorem 2), the Alice-Bob outage probability for a target rate (Theorem 3), and a numerically evaluated optimization of the effective covert rate via the jamming power limit (Proposition 4, Table I). The central claim is that positive-rate covert communication is possible for any ε>0 with E[Pe,w] ≥ 1−ε as n→∞, and that mmWave operation outperforms conventional RF covert communication.

Significance. If the results are interpreted in the right operational setting, the paper provides a useful and fairly tractable framework for jammer-assisted covert communication with directional mmWave beams. The derivations are largely transparent, the channel model is standard, and Theorem 1 is exact and cleanly proved. The numerical behavior in Figures 1 and 2 is consistent with the monotonicity statements. However, two issues materially reduce the significance as stated: the null hypothesis in Section III still has Alice transmitting a jamming beam, so the paper does not actually analyze 'hiding the existence of transmission' as claimed in the abstract; and the closed forms in Theorems 2 and 3 rely on Alzer's approximation used with equality signs, making those results approximate rather than exact. The framework is a legitimate study of jamming-assisted covertness, but the paper's framing and the exactness of its formulas need substantial correction.

major comments (2)
  1. [Section III, Eqs. (5)–(8); abstract and Section I] The covertness analysis uses a null hypothesis under which Alice is still transmitting. In Eq. (5), T_H0^w = PJ Gaw,s Law |h~_aw,s|^2 + σ_w^2, so under H0 Alice's second array radiates the jamming beam toward Willie. If Willie's task is to detect the existence of any transmission from Alice, as stated in the abstract and in the introduction's definition of covert communication, the correct H0 is T_H0^w = σ_w^2. In that case, for any Pa>0 the H1 energy exceeds the noise power almost surely, and Willie can choose τ slightly above σ_w^2 to obtain PFA=0 and PMD→0 as n→∞, i.e., P*e,w=0. The positive-rate result in Theorems 1–3 therefore establishes covertness of the data beam on top of a known, always-on jamming beam, not covertness of Alice's transmission. This is a legitimate jammer-assisted covertness model, but it is a structural modeling choice that must be stated prominently; as written, the abstract and conclusion go beyond what the model supports.
  2. [Theorems 2 and 3, Eqs. (16) and (24)] Alzer's lemma is an approximation, not an identity, yet the proofs use it with equality signs. In Eq. (16), step (a) states that Pr(X<x) 'can tightly be approximated' with [1−exp(−η_B x)]^{ν_B}, and the derivation then proceeds with '='; the same occurs in Eq. (24), step (b). Consequently Eqs. (11) and (21) are approximate closed forms, not exact characterizations. The paper should state this explicitly and, if possible, provide a bound on the approximation error. This matters for Proposition 4, which solves E[P*e,w] = 1−ε exactly to obtain the optimal Pmax_J; the equality is only approximate under the current derivation.
minor comments (4)
  1. [Proposition 4, Section IV-B] The statement 'P*AB_out is defined as (11)' refers to Eq. (11), which is the expression for E[P*e,w] from Theorem 2, not the outage probability. The intended meaning is presumably that P*AB_out is Theorem 3 evaluated at the Pmax_J,opt obtained from E[P*e,w] = 1−ε; this typo should be corrected.
  2. [Section III-B, Remark 2; Section IV-A] The closed-form results exclude the cases where Willie lies in the main lobe of Alice's first array and where Bob lies in the main lobe of the jamming array; Remark 2 only notes that additional averaging would be needed. This limitation should be stated in the abstract or conclusion, because the claimed superiority over RF is only demonstrated for the side-lobe configuration.
  3. [Table I and Section V] The effective rates R*_a,b in Table I are in bits per channel use, while the conclusion refers to 'much higher data rates, in bits per second' due to mmWave bandwidth. That extrapolation is not quantified and should be separated from the technical rate results.
  4. [Throughout] Several subscripted and superscripted symbols, such as g(a,s)_k and b(a,s)_k, are difficult to parse in the typeset equations; a table of notation would improve readability. Figures 1 and 2 also lose subscripts in the legends (e.g., 'a,s= 15o' should be θ_a,s = 15°).

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity found: all claimed results follow from the stated channel model, the uniform PJ assumption, and standard external lemmas; the H0 modeling choice is a scope limitation, not a circular step.

full rationale

The paper's derivation chain is self-contained and does not reduce any claimed result to its own inputs. Willie's detection error (Theorem 1) is computed directly from the two hypotheses in Eqs. (5) and (6), which are consequences of the stated received-signal model (2) and the independent uniform distribution of PJ; the optimal threshold interval and P*e,w then follow by elementary probability. Theorem 2 only averages this P*e,w over the channel fading, beamsteering, and LOS/NLOS states using Alzer's lemma and moment generating functions, with no fitted parameter renamed as a prediction. Theorem 3 derives the Alice-Bob outage probability from the SINR in Eq. (20), again using Alzer's lemma and the MGF of the gamma distribution. Proposition 4 numerically solves the equation E[P*e,w] = 1 - epsilon for Pmax_J, which is an optimization over a model parameter, not a fit to data used to manufacture a result. The only self-citations are [20] as motivational context and [24] for a routine integration technique in the proof of Theorem 2; neither is load-bearing or imported as an unverified uniqueness theorem. The reader's concern about H0 is a legitimate modeling critique: Eq. (5) includes the always-on jamming beam PJ G_aw,s L_aw |h_aw,s|^2, so 'hiding the existence of transmission' refers to the data beam on top of a known jamming beam rather than Alice's complete silence. However, this is a deliberate system-model assumption (stated in Section II-B and Section III), not a case where a conclusion is assumed by definition or where a fitted value is relabeled as a prediction. Therefore the circularity score is 0.

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

No parameters are fitted to data; numerical values such as path loss exponents, gains, and noise powers are illustrative examples from prior work. The central derivations rest on the stated channel model, the uniform jamming-power distribution, and standard probability lemmas, with the main caveat being the Alzer approximation treated as equality.

assumptions (6)
  • standard math Alzer's lemma approximation for the Gamma CDF is valid and tight for integer Nakagami parameters.
    Used in Theorems 2 and 3 to evaluate probabilities of ratios of Gamma random variables. It is an approximation, not an equality, which is a source of error not reflected in theorem statements.
  • domain assumption The jamming power P_J is uniformly distributed over [0, P_max_J] and is independent across blocks, with Willie knowing only the distribution.
    This uncertainty is what generates positive detection error; without it, Willie can separate H0 and H1 perfectly when the data signal is nonzero.
  • domain assumption The jamming beam is active in both H0 and H1, so Willie's task is to detect the data beam, not the transmitter's activity.
    Eq. (5) includes the P_J term under H0. This is necessary for the detection error analysis but conflicts with the abstract's framing of hiding the existence of transmission.
  • domain assumption Millimeter-wave channel model with LOS/NLOS path loss, deterministic blockage probability, sectored-pattern antenna gains with symmetric beamsteering errors, and Nakagami fading.
    The entire performance analysis is built on this model from references [16]-[19]; different channel models would change all expressions.
  • standard math Infinite blocklength and the strong law of large numbers make Willie's test statistic equal to the average received power.
    Used to write Tw in Eqs. (5)-(6) without finite-n fluctuations. Covertness is assessed in the n-to-infinity limit.
  • domain assumption The signals x_a and x_J have zero mean and unit power and are independent of the noise and each other.
    Needed for cross terms to vanish in Eqs. (5)-(6). The distribution beyond second moments is not specified, which limits the optimality claim for the energy detector.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Covert Millimeter-Wave Communication via a Dual-Beam Transmitter." pith.science (2026). https://pith.science/paper/RO6ZOZZK

@misc{pith2026190807591,
  author       = {Pith},
  title        = {Pith review of: Covert Millimeter-Wave Communication via a Dual-Beam Transmitter},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/RO6ZOZZK}},
  note         = {Machine review of arXiv:1908.07591}
}
read the original abstract

In this paper, we investigate covert communication over millimeter-wave (mmWave) frequencies. In particular, a dual-beam mmWave transmitter, comprised of two independent antenna arrays, attempts to reliably communicate to a receiver Bob when hiding the existence of transmission from a warden Willie. In this regard, operating over mmWave bands not only increases the covertness thanks to directional beams, but also increases the transmission data rates given much more available bandwidths and enables ultra-low form factor transceivers due to the lower wavelengths used compared to the conventional radio frequency (RF) counterpart. We assume that the transmitter Alice employs one of its antenna arrays to form a directive beam for transmission to Bob. The other antenna array is used by Alice to generate another beam toward Willie as a jamming signal with its transmit power changing independently from a transmission block to another block. We characterize Willie's detection performance with the optimal detector and the closed-form of its expected value from Alice's perspective. We further derive the closed-form expression for the outage probability of the Alice-Bob link, which enables characterizing the optimal covert rate that can be achieved using the proposed setup. Our results demonstrate the superiority of mmWave covert communication, in terms of covertness and rate, compared to the RF counterpart.

Figures

Figures reproduced from arXiv: 1908.07591 by the authors.

Figure 1
Figure 1. The expected value E[P ∗ e,w] of Willie’s detection error rate for a benchmark scenario with Ma,s = 15 dB, ma,f = −5 dB, Pa = 20 dBm, θa,s = 30o , and ∆ = 5o . The effect of different parameters is explored by considering the values Ma,s = 20 dB, ma,f = 0 dB, Pa = 5 dBm, θa,s = 15o , and ∆ = 15o while the rest of the parameters are exactly the same as the benchmark scenario. 0 5 10 15 20 25 30 35 40 45 50 P J max [d… view at source ↗
Figure 2
Figure 2. The outage probability of the Alice-Bob link for var [PITH_FULL_IMAGE:figures/full_fig_p006_2.png] view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

24 extracted references · 22 canonical work pages

  1. [1]

    Limits of reliabl e communi- cation with low probability of detection on AWGN channels,

    B. A. Bash, D. Goeckel, and D. Towsley, “Limits of reliabl e communi- cation with low probability of detection on AWGN channels,” IEEE J. Sel. Areas Commun. , vol. 31, no. 9, pp. 1921–1930, 2013

  2. [2]

    Hiding in formation in noise: fundamental limits of covert wireless communicat ion,

    B. A. Bash, D. Goeckel, D. Towsley, and S. Guha, “Hiding in formation in noise: fundamental limits of covert wireless communicat ion,” IEEE Commun. Mag. , vol. 53, no. 12, pp. 26–31, 2015

  3. [3]

    Reliable deniable com munication: Hiding messages in noise,

    P . H. Che, M. Bakshi, and S. Jaggi, “Reliable deniable com munication: Hiding messages in noise,” in IEEE Int. Symp. Inf. Theory (ISIT) , 2013, pp. 2945–2949

  4. [4]

    Fundamental limits of com- munication with low probability of detection,

    L. Wang, G. W. Wornell, and L. Zheng, “Fundamental limits of com- munication with low probability of detection,” IEEE Trans. Inf. Theory , vol. 62, no. 6, pp. 3493–3503, 2016

  5. [5]

    Covert communication over noisy channels: A resolvability perspective,

    M. R. Bloch, “Covert communication over noisy channels: A resolvability perspective,” IEEE Trans. Inf. Theory , vol. 62, pp. 2334–2354, 2016

  6. [6]

    First-and second-order asy mptotics in covert communication,

    M. Tahmasbi and M. R. Bloch, “First-and second-order asy mptotics in covert communication,” IEEE Transactions on Information Theory , vol. 65, no. 4, pp. 2190–2212, 2018

  7. [7]

    Embedding covert infor mation in broadcast communications,

    K. S. K. Arumugam and M. R. Bloch, “Embedding covert infor mation in broadcast communications,” IEEE Transactions on Information F orensics and Security , vol. 14, no. 10, pp. 2787–2801, 2019

  8. [8]

    Covert Communication over a K-User Multiple Access Channel

    ——, “Covert communication over a k-user multiple access channel,” arXiv preprint arXiv:1803.06007 , 2018

Show all 24 references
  1. [9]

    Achieving undetectable communication,

    S. Lee, R. J. Baxley, M. A. Weitnauer, and B. Walkenhorst, “Achieving undetectable communication,” IEEE J. Sel. Topics Signal Process. , vol. 9, no. 7, pp. 1195–1205, 2015

  2. [10]

    Covert com munications when the warden does not know the background noise power,

    D. Goeckel, B. Bash, S. Guha, and D. Towsley, “Covert com munications when the warden does not know the background noise power,” IEEE Commun. Lett. , vol. 20, no. 2, pp. 236–239, 2016

  3. [11]

    Covert communic ation gains from adversary’s ignorance of transmission time,

    B. A. Bash, D. Goeckel, and D. Towsley, “Covert communic ation gains from adversary’s ignorance of transmission time,” IEEE Trans. Wireless Commun., vol. 15, no. 12, pp. 8394–8405, 2016

  4. [12]

    Covert c ommunication achieved by a greedy relay in wireless networks,

    J. Hu, S. Y an, X. Zhou, F. Shu, J. Li, and J. Wang, “Covert c ommunication achieved by a greedy relay in wireless networks,” IEEE Trans. Wireless Commun., vol. 17, no. 7, pp. 4766–4779, 2018

  5. [13]

    Covert c ommu- nication with the help of relay and channel uncertainty,

    J. Wang, W. Tang, Q. Zhu, X. Li, H. Rao, and S. Li, “Covert c ommu- nication with the help of relay and channel uncertainty,” IEEE Wireless Commun. Lett. , vol. 8, no. 1, pp. 317–320, 2019

  6. [14]

    Covert communication i n fading channels under channel uncertainty,

    K. Shahzad, X. Zhou, and S. Y an, “Covert communication i n fading channels under channel uncertainty,” in IEEE VTC Spring . IEEE, 2017, pp. 1–5

  7. [15]

    Cove rt commu- nications with a full-duplex receiver over wireless fading channels,

    J. Hu, K. Shahzad, S. Y an, X. Zhou, F. Shu, and J. Li, “Cove rt commu- nications with a full-duplex receiver over wireless fading channels,” in IEEE Int. Conf. Commun. (ICC) , 2018, pp. 1–6

  8. [16]

    Millimeter wave mobile communications for 5G cellular: It will work!

    T. S. Rappaport, S. Sun, R. Mayzus, H. Zhao, Y . Azar, K. Wa ng, G. N. Wong, J. K. Schulz, M. Samimi, and F. Gutierrez, “Millimeter wave mobile communications for 5G cellular: It will work!” IEEE access , vol. 1, pp. 335–349, 2013

  9. [17]

    Modeling and analyzing millimeter wave cellul ar systems,

    J. G. Andrews, T. Bai, M. N. Kulkarni, A. Alkhateeb, A. K. Gupta, and R. W. Heath, “Modeling and analyzing millimeter wave cellul ar systems,” IEEE Trans. Commun. , vol. 65, no. 1, pp. 403–430, 2017

  10. [18]

    Coverage and rate analysis for mi llimeter-wave cellular networks,

    T. Bai and R. W. Heath, “Coverage and rate analysis for mi llimeter-wave cellular networks,” IEEE Trans. Wireless Commun. , vol. 14, no. 2, pp. 1100–1114, 2015

  11. [19]

    Stochastic geometry modeling and analysi s of multi- tier millimeter wave cellular networks,

    M. Di Renzo, “Stochastic geometry modeling and analysi s of multi- tier millimeter wave cellular networks,” IEEE Trans. Wireless Commun. , vol. 14, no. 9, pp. 5038–5057, 2015

  12. [20]

    Channel coding at low capacity,

    M. Fereydounian, M. V . Jamali, H. Hassani, and H. Mahdav ifar, “Channel coding at low capacity,” arXiv preprint arXiv:1811.04322 , 2018

  13. [21]

    Covert communication in the presence of an uninformed jammer,

    T. V . Sobers, B. A. Bash, S. Guha, D. Towsley, and D. Goeck el, “Covert communication in the presence of an uninformed jammer,” IEEE Trans. Wireless Commun., vol. 16, no. 9, pp. 6193–6206, 2017

  14. [22]

    I. S. Gradshteyn and I. M. Ryzhik, Table of integrals, series, and products. Academic Press, 2007

  15. [23]

    On some inequalities for the incomplete Gamm a function,

    H. Alzer, “On some inequalities for the incomplete Gamm a function,” Math. Comput. , vol. 66, no. 218, pp. 771–778, 1997

  16. [24]

    Uplink non-orthogonal multiple access over mixed RF-FSO systems,

    M. V . Jamali and H. Mahdavifar, “Uplink non-orthogonal multiple access over mixed RF-FSO systems,” arXiv preprint arXiv:1903.00326 , 2019

Pith tools

Reviewed August 14, 2026 · model on record in the stance chip above.