Pith. sign in

REVIEW 3 major objections 4 minor 47 references

Experimental Covert Communication Using Software-Defined Radio

T0 review · 3 major / 4 minor · reviewed 2026-08-07 · deepseek-v4-flash

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

Pith's one-line read A software-defined radio link transmits covertly at the square-root-law rate, with the warden's detection error near chance.

desk verdict First RF SRL covert-communication experiment with real SDR hardware—valuable proof-of-concept, but the covertness proof has a gap that needs referee attention. read the letter →

arxiv 2506.02297 v1 pith:OY5RQ55F submitted 2025-06-02 cs.NI cs.ITcs.SYeess.SYmath.IT

Rohan Bali , Trevor E. Bailey , Michael S. Bullock , Boulat A. Bash This is my paper
classification cs.NIcs.ITcs.SYeess.SYmath.IT
keywords squarerootlawcovertcommunicationlowprobabilityofdetectionsoftware-definedradioAWGNchannelsparsecodingbinaryphase-shiftkeyingwarden
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

This paper reports the first radio-frequency implementation of square-root-law covert communication, a regime in which a transmitter hides the very existence of a message, not just its content. Using a software-defined radio sender, receiver, and a warden radio that listens for any sign of transmission, the authors show that a sparse-coding scheme transmits covert bits at the predicted O(sqrt(n)) rate while the warden's optimal detector performs no better than chance. The receiver decodes symbols with error probability around 0.17, and the warden's detection error stays near the target 0.45, matching the theoretical bound. If correct, this turns a decade of information-theoretic results into a working blueprint for low-probability-of-detection radio links.

What carries the argument

The load-bearing mechanism is sparse coding with a secretly shared Bernoulli transmission pattern: each pulse slot is used with probability alpha_n ~ 1/$\sqrt$(n), and unused slots carry pure noise. Because the warden lacks the pattern, each slot's distribution is a mixture of noise and a BPSK signal, and the relative entropy between the 'on' and 'off' hypotheses is controlled by a Taylor-series bound that yields alpha_n = 2 $SNR^{{-1}}$ $\sqrt$(delta_RE / n_p). Gaussian pulse shaping confines the signal energy to 76-sample slots, and a single pilot symbol compensates the carrier phase for the receiver. The combination converts the square-root law from an asymptotic statement into a finite-system design rule whose predictions are then measured.

What would settle it

Compute the sixth-order term in the Taylor expansion of the relative entropy D(P0||P1) for the actual experimental parameters (pulse norm, SNR, alpha_n) and check whether it stays negative over [0, ||c||]; if it changes sign, the claimed upper bound on Willie's detection error is not established. Alternatively, directly estimate Willie's total variation distance from raw I/Q samples and see whether the measured detection error falls below the 0.45 target at any transmission duration.

Watch

Extended reading notes

Core claim

The paper's central claim is that the square root law, previously validated only on optical channels, holds on radio-frequency channels built from software-defined radios. The scheme fixes a pulse shape of norm ||c||, divides time into pulse slots, and transmits in each slot with probability alpha_n set to O(1/$\sqrt$(n)) according to a pre-shared random pattern. BPSK symbols are placed in the selected slots, and the warden, who knows all system parameters except the secret pattern, must decide between noise and transmission. An upper bound on the relative entropy between the noise-only and transmission distributions, derived from a Taylor expansion, yields the detection-error lower bound near 0.45; the experiment confirms this while Bob's decoding error stays near 0.17. The total number of covert bits scales as B_bsc(n) = O($\sqrt$(n)), with a log-log slope fit of one-half ($R^{2}$ = 0.96).

Load-bearing premise

The covertness proof assumes a Taylor-series remainder term in the relative-entropy expansion is negative over the whole range of pulse amplitudes used in the experiment; the paper says this was checked numerically but gives no details, so a reader cannot independently verify the detection-error bound.

Editorial extensions

If this is right

  • If the claim holds, other RF systems can adopt the same sparse-coding recipe: pre-shared Bernoulli pattern, Gaussian pulse shaping, pilot-based phase correction, and a transmission probability proportional to 1/sqrt(n).
  • Provably covert links become a hardware design target rather than a purely theoretical construction, at least for static AWGN-like channels.
  • The measured decoding error near 0.17, combined with the warden's near-chance detection, shows that the same channel can simultaneously support reliable decoding and covertness at the predicted scaling.
  • The observed R^2 = 0.96 for a one-half slope in the bit-count versus time plot is quantitative evidence that the square root law, not some other rate, governs this implementation.
  • Future systems can trade the conservative bounds for the larger constants available when the warden lacks timing, frequency, or noise-power knowledge, as the paper notes for its own follow-up.

Reading between the lines

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

  • The paper does not say this, but a natural extension is to relax the assumption of an always-on, spectrally flat noise source; using a real, varying noise floor would test whether the scheme degrades gracefully in the wild.
  • The same sparse-coding machinery could be adapted to other modulations, such as QPSK or OFDM, and the paper's explicit plan to try QPSK suggests the BPSK choice is not fundamental to the square-root-law behavior.
  • One could build a detector-side test that directly estimates the warden's likelihood ratio from the collected I/Q data; the paper plans this as future work, and it would provide a stronger empirical check than the current bound-based estimate.
  • The reliance on a pre-shared secret of size O(sqrt(n) log n) hints that the practical bottleneck for covert RF may shift from information-theoretic limits to key management and synchronization.
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

3 major / 4 minor

Summary. The manuscript presents an SDR-based implementation of square-root-law (SRL) covert communication over an RF AWGN channel. Alice and Bob share a random Bernoulli schedule t; Alice transmits BPSK pulses only in selected pulse slots with probability α_n=O(1/sqrt(n)), and Bob performs hard-decision decoding using the secret schedule. The authors derive α_n from a Taylor expansion of the relative entropy at Willie (Eq. (5)), estimate Willie's SNR from a calibration transmission, and run 500 trials per transmission duration on the ORBIT testbed with USRP X310 radios connected by coaxial cables and an artificial Gaussian noise source. They report Bob's raw symbol error probability p_e,bsc≈0.17, from which they compute an SRL-compatible number of reliable bits B_bsc(n) (Fig. 4), and they plot a lower bound on Willie's detection error probability near 0.45 (Fig. 5). The paper claims to demonstrate the first experimental validation of SRL-based covert communication in the RF domain.

Significance. If the claims were fully supported, this would be a valuable first RF demonstration of a theoretical covertness law that has only been validated optically. The experimental apparatus is carefully described: synchronized USRPs, a controlled wired channel, large N, and rolling-buffer data processing. The stable raw error rate near 0.17 and stable warden SNR across time scales are meaningful feasibility data. However, the significance is tempered by three gaps: the covertness proof relies on an unverified and numerically contradicted Taylor-remainder assumption; the sqrt(n) scaling of B_bsc is imposed by the design equations rather than measured independently; and no error-correcting code was run, so 'reliable bits' are capacity estimates. These issues make the current version an interesting feasibility study rather than a conclusive validation.

major comments (3)
  1. [Appendix A / Remark 1 / Eq. (5) / Fig. 5] The δ-covertness guarantee D(P0^n||P1^n) ≤ δRE is not established for the parameters used. The proof relies on the assertion, stated in Remark 1 without details, that the sixth-order Taylor remainder is negative for every ξ in [0, ||c||]. A direct calculation for the reported operating point (T = 0.5 s, SNR ≈ 1.6, α_n ≈ 6.2×10^-4, np ≈ 8.2×10^4) gives D_slot = E[-log(1 + α_n(r(Z)-1))] with r(Z) = exp(-SNR/2) cosh(√SNR Z). To leading order in α_n, D_slot ≈ (α_n^2/2)(cosh(SNR)-1) ≈ 3.0×10^-7, so D_total ≈ 0.025 > δRE = 0.02. Thus the sixth-order remainder is positive at this operating point, and the plotted lower bound p_e^(w) ≈ 0.45 in Fig. 5 is not a proven lower bound; the Pinsker bound with the correct constant gives p_e^(w) ≥ 0.444 (or 0.39 under the paper's missing ln2 convention). The authors must either report the promised numerical verification, recompute α_n so that the exact relative entropy satisfies the bound, or replace the claim of provable δ-covertness with an explicitly weaker statement.
  2. [Sec. III-D / Fig. 4 / Eqs. (2), (5)] The SRL-scaling evidence in Fig. 4 is largely by construction. B_bsc(n) is computed as α_n np C_bsc via Eq. (2), with α_n set by Eq. (5) to 2 SNR^{-1} sqrt(δRE/np). Since α_n already contains the 1/sqrt(np) factor, B_bsc(n) necessarily grows as sqrt(n) for constant C_bsc; the R^2 = 0.96 fit with slope 1/2 therefore does not constitute an independent empirical test of the square root law. In addition, no error-correcting code was actually run; the 'reliable bits' are capacity estimates based on the measured raw error p_e,bsc ≈ 0.17, not decoded bits. The experimental contribution is better described as demonstrating that the hardware can sustain the sparse-coding scheme with stable raw error and stable warden SNR, not as a direct validation of SRL scaling.
  3. [Sec. II-C / Eq. (4)] There is a factor error in the covertness criterion. Equation (4) states p_e ≥ 1/2 - (1/2) sqrt(2D), with D defined using log2 in the text. Pinsker's inequality with base-2 relative entropy is p_e ≥ 1/2 - (1/2) sqrt((ln2/2)D); the form sqrt(2D) is correct only if D denotes the natural-log relative entropy. As written, D ≤ δRE = 8δ^2 with δ = 0.05 gives p_e ≥ 0.4, not the 0.45 target. Please make the base of the logarithm explicit and carry the ln2 factor (or define D with ln) consistently through Eq. (5) and the Appendix.
minor comments (4)
  1. [Sec. III-C vs. Figs. 4-5] The text says 'T ∈ [0.05, 15] s' but the figures show ten values from 0.50 s to 5.00 s; please align the stated range with the plotted values.
  2. [Fig. 5 caption / Sec. III-D] The left ordinate of Fig. 5 is a theoretical lower bound derived from the relative-entropy upper bound, not a measured detection error probability; please state this explicitly in the caption and axis label, since the current label 'Detection error probability p_e^(w)' may mislead readers.
  3. [Sec. III-C] The sentence 'we compute α_n√n = 4√2δ/SNR' uses n ambiguously: Eq. (5) and the implementation require n to be the number of pulse slots np, not the number of channel uses n = fsT. Please disambiguate the notation in the experimental section.
  4. [Appendix B] Appendix B says the phase estimate is obtained by averaging over many instances of p_I and p_Q, while the system description says a single pilot symbol is sufficient; please clarify how many pilot symbols were actually transmitted and processed.

Circularity Check

3 steps flagged · score 6.0 of 10

SRL scaling and Willie's covertness bound are built into the experiment's parameter choice; the genuinely measured quantities do not test the central prediction.

  1. fitted input called prediction [Section III-C, Section III-D, Eq. (2), Eq. (5), Fig. 4]
    "We use the SNR estimate to compute αn√n = 4√2δ SNR. We set δ = 0.05 and use the result to compute αn for each value of T. ... We observe that p(b)e,bsc ≈ 0.17 throughout our experiments. Using the left ordinate we report the corresponding estimate of the total number of transmissible covert bits Bbsc(n) using the equality in (2). Fitting a line with slope of one-half to the log-log plot results in the coefficient of determination R2 = 0.96, indicating the SRL-scaling that we expect."

    In Section III-C, alpha_n is set to 2 SNR^{-1} sqrt(delta_RE/n_p) (Eq. (5)), i.e., alpha_n = O(1/sqrt(n_p)), before any data are collected. Eq. (2) then defines B_bsc(n) ≈ alpha_n n_p C_bsc. Substituting Eq. (5) gives B_bsc(n) = (2 C_bsc sqrt(delta_RE)/(SNR sqrt(n_s))) sqrt(n), so the log-log slope of B_bsc versus T is exactly 1/2 by construction, independent of the measurements. The R^2 = 0.96 line therefore fits a constructed quantity, not an empirical test of square-root-law scaling; the only data-dependent factor is the roughly constant C_bsc, and the observed p_e,bsc ≈ 0.17 was not used to set alpha_n.

  2. self definitional [Section III-C, Section III-D, Eq. (5), Fig. 5]
    "We set δ = 0.05 and use the result to compute αn for each value of T. ... We then estimate the lower bound on Willie's probability of error p(w)e by computing the upper bound on relative entropy derived from Taylor series expansion in Appendix A (and verifying that it is indeed an upper bound per Remark 1 therein). We plot it using the left ordinate and note that it is very conservative, as it is, effectively, a lower bound on a lower bound. Nevertheless, this is sufficient to show that we indeed achieve covert communication."

    The plotted p_e^(w) lower bound is not a measured detection-error probability; it is the Pinsker bound computed from the Taylor upper bound on D(P_0^n||P_1^n). The same delta_RE and Eq. (5) were used to choose alpha_n, so the bound p_e >= 0.5 - 0.5 sqrt(2 delta_RE) is fixed by the design input before the experiment. Re-evaluating that bound with the measured SNR and plotting it near the 'target' line cannot confirm covertness empirically; it restates the construction used to set alpha_n. A real validation would estimate Willie's detection-error rate from his observed samples, which the paper defers to future work.

1 more flagged steps
  1. other [Appendix A, Remark 1]
    "The argument using Taylor's theorem with remainder from [47, Th. 1] is contingent on the BPSK signal power being arbitrarily small. We may adapt this argument for our experimental setup by showing that the sixth term in the expansion is negative for every ξ ∈ [0, ∥c∥]. We verify this numerically for our estimated experimental parameters but omit the details for brevity."

    This is the load-bearing step that extends the small-power Taylor result [47] (by the same authors) to the finite pulse amplitude of the experiment. The condition that the sixth-order remainder be negative on [0, ||c||] is asserted to be verified numerically, but the verification is omitted. If the condition fails, D(P_0^n||P_1^n) <= delta_RE, and hence the claimed p_e^(w) >= target, is not established. As written, the covertness proof reduces to an unshown numerical check plus a self-citation, so the paper's central guarantee is not independently verifiable from the text.

full rationale

The paper contains genuine experimental content: the SDR implementation, the measured Bob-side BPSK error probability p_e,bsc ≈ 0.17, the stable Willie-side SNR estimates, and the control of the ORBIT testbed are real empirical contributions. However, the two headline validation metrics are not independent tests of the theory. The SRL scaling shown in Fig. 4 is forced because alpha_n is set by Eq. (5) to be proportional to 1/sqrt(n_p), and B_bsc(n) is then computed from Eq. (2) using that same alpha_n; the R^2 = 0.96 slope-1/2 fit is a fit to a constructed curve. Likewise, Willie's reported detection-error 'performance' in Fig. 5 is not a measured error probability but a lower bound computed from the same Taylor-series relative-entropy bound that was used to choose alpha_n, so it cannot fail and does not empirically validate covertness. The load-bearing Taylor-remainder step in Appendix A is imported from the authors' earlier work [47] and depends on an unshown numerical verification (Remark 1), so the central p_e^(w) guarantee is not established from the printed text. These reductions make the paper partially circular: the claimed validation of the square-root law and of Willie's covertness reduces to construction, even though the hardware measurements themselves are not fabricated. Score 6 rather than higher because the measured Bob error rate and SNR stability are independent data and are not used to set alpha_n.

Assumptions & free parameters 3 free parameters · 6 assumptions · 0 invented entities

The central claim rests on standard information-theoretic inequalities, a Taylor-expansion covertness argument whose remainder term is asserted but not shown, and a threat model in which Willie knows all system parameters except the secret schedule. No new physical entities are introduced; the sparse-coding and one-time-pad components are protocol constructs.

free parameters (3)
  • Target detection error delta = 0.05
    Chosen by the authors; sets delta_RE = 8 delta^2 in Eq. (5) and hence the covert bit count. Different delta changes B_bsc and the required alpha_n.
  • Pulse-slot length n_s = 76 samples
    Implementation choice that enters the SRL constant through B_bsc proportional to sqrt(n_p) / sqrt(n_s). It affects the number of covert bits but not the sqrt scaling.
  • Willie's received SNR estimate = about 2 to 3 dB (initial calibration)
    Estimated once in a calibration run and used to set alpha_n for all trials. The covertness bound and B_bsc scale inversely with this SNR; the paper checks that it stays roughly constant over time.
assumptions (6)
  • standard math Pinsker's inequality and the total variation lower bound on detection error (Eq. (4))
    Used to convert relative entropy into a bound on Willie's error probability.
  • standard math Taylor's theorem with remainder for the relative entropy expansion at ||c||=0, following [47]
    Basis for Eq. (5); the fourth-order term is computed and assumed dominant.
  • ad hoc to paper The sixth term in the Taylor expansion is negative for every xi in [0, ||c||] (Remark 1)
    This is load-bearing for the upper bound on relative entropy at finite pulse amplitude; verified numerically but not shown.
  • domain assumption Willie knows transmission start time, channel conditions, pulse shape, and alpha, but not the secret schedule t or one-time pad s
    Standard SRL threat model; the experiment implements this by giving Willie matched knowledge except the secrets.
  • domain assumption The channel is discrete-time AWGN with static path loss and phase; noise is circularly symmetric Gaussian
    The theoretical model in Section II-A; realized with a wired star network and a synthetic Gaussian noise source.
  • domain assumption The pre-shared secret is generated fresh per trial and the one-time pad makes transmitted symbols equiprobable
    Needed so Willie's H1 distribution is the mixture in Eq. (7).

how reviews work

0 comments
Cite this review

Pith. "Pith review of Experimental Covert Communication Using Software-Defined Radio." pith.science (2026). https://pith.science/paper/OY5RQ55F

@misc{pith2026250602297,
  author       = {Pith},
  title        = {Pith review of: Experimental Covert Communication Using Software-Defined Radio},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/OY5RQ55F}},
  note         = {Machine review of arXiv:2506.02297}
}
abstract

The fundamental information-theoretic limits of covert, or low probability of detection (LPD), communication have been extensively studied for over a decade, resulting in the square root law (SRL): only $L\sqrt{n}$ covert bits can be reliably transmitted over time-bandwidth product $n$, for constant $L>0$. Transmitting more either results in detection or decoding errors. The SRL imposes significant constraints on hardware realization of provably-secure covert communication. Thus, experimental validation of covert communication is underexplored: to date, only two experimental studies of SRL-based covert communication are available, both focusing on optical channels. Here, we report our initial results demonstrating the provably-secure covert radio-frequency (RF) communication using software-defined radios (SDRs). These validate theoretical predictions, open practical avenues for implementing covert communication systems, as well as raise future research questions.

Figures

Figures reproduced from arXiv: 2506.02297 by the authors.

Figure 1
Figure 1. Covert communication over an AWGN channel. Alice ( [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. Covert communication experiment on ORBIT. Four Ettus USRP X310 [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 3
Figure 3. Data-packet structure (durations in samples). BPSK modulates one bit per symbol using 76 samples: a sample modulating the bit followed by the [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗
Figures from the paper (2 more)
Figure 4
Figure 4. Figure 4: Bob’s receiver performance. Total number [PITH_FULL_IMAGE:figures/full_fig_p005_4.png]
Figure 5
Figure 5. Figure 5: Willie’s detector performance. Detection probability error [PITH_FULL_IMAGE:figures/full_fig_p005_5.png]

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

47 extracted references · 43 canonical work pages

  1. [1]

    Square root law for communication with low probability of detection on AWGN channels,

    B. A. Bash, D. Goeckel, and D. Towsley, “Square root law for communication with low probability of detection on AWGN channels,” in Proc. IEEE Int. Symp. Inform. Theory (ISIT) , Cambridge, MA, Jul. 2012

    2012

  2. [2]

    Limits of reliable communication with low probability of detection on AWGN channels,

    ——, “Limits of reliable communication with low probability of detection on AWGN channels,”IEEE J. Select. Areas Commun., vol. 31, no. 9, pp. 1921–1930, 2013

    work page 1921

  3. [3]

    Hiding information in noise: Fundamental limits of covert wireless communication,

    B. A. Bash, D. Goeckel, S. Guha, and D. Towsley, “Hiding information in noise: Fundamental limits of covert wireless communication,” IEEE Commun. Mag., vol. 53, no. 12, 2015

    work page 2015

  4. [4]

    Covert communications: A comprehensive survey,

    X. Chen, J. An, Z. Xiong, C. Xing, N. Zhao, F. R. Yu, and A. Nal- lanathan, “Covert communications: A comprehensive survey,” IEEE Commun. Surv. Tutor., vol. 25, no. 2, pp. 1173–1198, 2023

    work page 2023

  5. [5]

    A. J. Menezes, S. A. Vanstone, and P. C. V . Oorschot, Handbook of Applied Cryptography. Boca Raton, FL, USA: CRC Press, Inc., 1996

    work page 1996

  6. [6]

    M. R. Bloch and J. Barros, Physical-Layer Security . Cambridge University Press, 2011

    work page 2011

  7. [7]

    M. K. Simon, J. K. Omura, R. A. Scholtz, and B. K. Levitt, Spread Spectrum Communications Handbook . McGraw-Hill, 1994

    work page 1994

  8. [8]

    T. M. Cover and J. A. Thomas, Elements of Information Theory, 2nd ed. John Wiley & Sons, Hoboken, NJ, 2002

    work page 2002

Show all 47 references
  1. [9]

    Covert communication over noisy channels: A resolvabil- ity perspective,

    M. R. Bloch, “Covert communication over noisy channels: A resolvabil- ity perspective,” IEEE Trans. Inf. Theory, vol. 62, no. 5, pp. 2334–2354, May 2016

    2016

  2. [10]

    Fundamental limits of communication with low probability of detection,

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

    2016

  3. [11]

    First and second order asymptotics in covert communication,

    M. Tahmasbi and M. R. Bloch, “First and second order asymptotics in covert communication,” IEEE Trans. Inf. Theory , vol. 65, no. 4, pp. 2190–2212, Apr. 2019

    2019

  4. [12]

    On the second order asymptotics of covert communications over awgn channels,

    Y . Xinchun, S. Wei, S.-L. Huang, and X. P. Zhang, “On the second order asymptotics of covert communications over awgn channels,” in Proc. IEEE Int. Conf. Commun. (ICC) , 2024, pp. 1479–1484

    2024

  5. [13]

    Covert communication over broadcast channels,

    K. S. K. Arumugam and M. R. Bloch, “Covert communication over broadcast channels,” in Proc. Inform. Theory Workshop (ITW) , Nov. 2017, pp. 299–303

    2017

  6. [14]

    Covert communication in wireless relay networks,

    J. Hu, S. Yan, X. Zhou, F. Shu, and J. Wang, “Covert communication in wireless relay networks,” arXiv:1704.04946 [cs.IT], Apr. 2017

    2017 arXiv

  7. [15]

    Covert communication over a k-user multiple access channel,

    K. S. K. Arumugam and M. R. Bloch, “Covert communication over a k-user multiple access channel,” IEEE Trans. Inf. Theory , vol. 65, no. 11, pp. 7020–7044, Nov. 2019

    2019

  8. [16]

    Time-division is optimal for covert com- munication over some broadcast channels,

    V . Y . F. Tan and S. Lee, “Time-division is optimal for covert com- munication over some broadcast channels,” IEEE Trans. Inf. Forensics Security, vol. 14, no. 5, pp. 1377–1389, May 2019

    2019

  9. [17]

    Multi-hop routing in covert wireless networks,

    A. Sheikholeslami, M. Ghaderi, D. Towsley, B. A. Bash, S. Guha, and D. Goeckel, “Multi-hop routing in covert wireless networks,” IEEE Trans. Wireless Commun., vol. 17, no. 6, pp. 3656–3669, Jun. 2018

    2018

  10. [18]

    Covert wireless communication with artificial noise generation,

    R. Soltani, D. Goeckel, D. Towsley, B. A. Bash, and S. Guha, “Covert wireless communication with artificial noise generation,” IEEE Trans. Wireless Commun., vol. 17, no. 11, pp. 7252–7267, Nov. 2018

    2018

  11. [19]

    Quantum Noise Limited Communication with Low Probability of Detection,

    B. A. Bash, S. Guha, D. Goeckel, and D. Towsley, “Quantum Noise Limited Communication with Low Probability of Detection,” in Proc. IEEE Int. Symp. Inform. Theory (ISIT) , Istanbul, Turkey, Jul. 2013

    2013

  12. [20]

    Quantum-secure covert communication on bosonic channels,

    B. A. Bash, A. H. Gheorghe, M. Patel, J. L. Habif, D. Goeckel, D. Towsley, and S. Guha, “Quantum-secure covert communication on bosonic channels,” Nat. Commun., vol. 6, Oct. 2015

    2015

  13. [21]

    Covert communication over classical-quantum channels,

    A. Sheikholeslami, B. A. Bash, D. Towsley, D. Goeckel, and S. Guha, “Covert communication over classical-quantum channels,” in Proc. IEEE Int. Symp. Inform. Theory (ISIT) , Barcelona, Spain, Jul. 2016

    2016

  14. [22]

    Fundamental limits of quantum-secure covert communication over bosonic channels,

    M. S. Bullock, C. N. Gagatsos, S. Guha, and B. A. Bash, “Fundamental limits of quantum-secure covert communication over bosonic channels,” IEEE J. Sel. Areas Commun. , vol. 38, no. 3, pp. 471–482, Mar. 2020

    2020

  15. [23]

    Covert capacity of bosonic channels,

    C. N. Gagatsos, M. S. Bullock, and B. A. Bash, “Covert capacity of bosonic channels,” IEEE J. Sel. Areas Inf. Theory , vol. 1, pp. 555–567, 2020

    2020

  16. [24]

    Toward undetectable quantum key distribution over bosonic channels,

    M. Tahmasbi and M. R. Bloch, “Toward undetectable quantum key distribution over bosonic channels,” arXiv:1904.12363 [cs.IT], 2019

    1904 arXiv

  17. [25]

    Toward undetectable quantum key distribution over bosonic channels,

    ——, “Toward undetectable quantum key distribution over bosonic channels,” IEEE J. Sel. Areas Inf. Theory , vol. 1, no. 2, pp. 585–598, 2020

    2020

  18. [26]

    Fundamental limits of bosonic broadcast channels,

    E. J. Anderson, S. Guha, and B. A. Bash, “Fundamental limits of bosonic broadcast channels,” in Proc. IEEE Int. Symp. Inform. Theory (ISIT), virtual, Jul. 2021

    2021

  19. [27]

    Towards a characterization of the covert capacity of bosonic channels under trace distance,

    S.-Y . Wang, T. Erdo˘gan, and M. Bloch, “Towards a characterization of the covert capacity of bosonic channels under trace distance,” in Proc. IEEE Int. Symp. Inf. Theory (ISIT) . IEEE Press, 2022, pp. 318–323

    2022

  20. [28]

    Square root law for covert quantum communication over optical channels,

    E. J. D. Anderson, C. K. Eyre, I. M. Dailey, F. Rozp˛ edek, and B. A. Bash, “Square root law for covert quantum communication over optical channels,” in Proc. IEEE Int. Conf. Quantum Comput. Eng. (QCE) , Montréal, QC, Canada, 2024

    2024

  21. [29]

    Fundamental limits of covert communication over classical-quantum channels,

    M. S. Bullock, A. Sheikholeslami, M. Tahmasbi, R. C. Macdonald, S. Guha, and B. A. Bash, “Fundamental limits of covert communication over classical-quantum channels,” 2025, to appear in IEEE Trans. Inf. Theory. [Online]. Available: https://arxiv.org/abs/1601.06826

    2025 arXiv

  22. [30]

    Entanglement-assisted covert communication via qubit depolarizing channels,

    E. Zlotnick, B. A. Bash, and U. Pereg, “Entanglement-assisted covert communication via qubit depolarizing channels,” IEEE Trans. Inf. Theory, vol. 71, no. 5, pp. 3693–3706, 2025

    2025

  23. [31]

    Experimental covert communication over metropolitan fibre optical links,

    Y . Liu, J. M. Arrazola, W.-Z. Liu, W. Zhang, I. W. Primaatmaja, H. Li, L. You, Z. Wang, Q. Zhang, and J.-W. Pan, “Experimental covert communication over metropolitan fibre optical links,” IEEE Wireless Commun., vol. 31, no. 4, pp. 76–80, 2024

    2024

  24. [32]

    Overview of the ORBIT radio grid testbed for evaluation of next-generation wireless network protocols,

    D. Raychaudhuri, I. Seskar, M. Ott, S. Ganu, K. Ramachandran, H. Kremo, R. Siracusa, H. Liu, and M. Singh, “Overview of the ORBIT radio grid testbed for evaluation of next-generation wireless network protocols,” in IEEE Wireless Commun. Netw. Conf. (WCNC) , vol. 3, 2005, pp. 1664–1669

    2005

  25. [33]

    Grid domain overview,

    ORBIT Project, “Grid domain overview,” https://www.orbit-lab.org/ wiki/Hardware/bDomains/aGrid, 2025, accessed: April 16, 2025

    2025

  26. [34]

    LPD Communication when the Warden Does Not Know When,

    B. A. Bash, D. Goeckel, and D. Towsley, “LPD Communication when the Warden Does Not Know When,” in Proc. IEEE Int. Symp. In- form. Theory (ISIT) , Honolulu, HI, Jul. 2014

    2014

  27. [35]

    Covert communication gains from adversary’s ignorance of transmission time,

    ——, “Covert communication gains from adversary’s ignorance of transmission time,” IEEE Transactions on Wireless Communications , vol. 15, no. 12, pp. 8394–8405, Dec. 2016, Originally presented at ISIT 2014, Honolulu HI

    2016

  28. [36]

    Keyless asynchronous covert communication,

    K. S. K. Arumugam and M. R. Bloch, “Keyless asynchronous covert communication,” in Proc. Inform. Theory Workshop (ITW) , Sep. 2016

    2016

  29. [37]

    Covert communication with the help of an uninformed jammer achieves positive rate,

    T. V . Sobers, B. A. Bash, D. Goeckel, S. Guha, and D. Towsley, “Covert communication with the help of an uninformed jammer achieves positive rate,” in Asilomar Conf. Signals Syst. Comput. , Pacific Grove, CA, USA, Nov. 2015, pp. 625–629

    2015

  30. [38]

    Covert communication in the presence of an uninformed jammer,

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

    2017

  31. [39]

    Explicit design of provably covert channel codes,

    S.-Y . Wang and M. R. Bloch, “Explicit design of provably covert channel codes,” in IEEE Int. Symp. Inf. Theory (ISIT) , 2021, pp. 190– 195

    2021

  32. [40]

    Richardson and R

    T. Richardson and R. Urbanke, Modern Coding Theory . Cambridge University Press, 2008

    2008

  33. [41]

    Fridrich, Steganography in Digital Media: Principles, Algorithms, and Applications , 1st ed

    J. Fridrich, Steganography in Digital Media: Principles, Algorithms, and Applications , 1st ed. New York: Cambridge University Press, 2009

    2009

  34. [42]

    Lehmann and J

    E. Lehmann and J. Romano, Testing Statistical Hypotheses , 3rd ed. New York: Springer, 2005

    2005

  35. [43]

    Octoclock and octoclock-g product overview,

    Ettus Research, “Octoclock and octoclock-g product overview,” https://www.ettus.com/wp-content/uploads/2019/01/Octoclock\_Spec\ _Sheet.pdf, Jan. 2019, accessed: Apr. 9, 2025

    2019

  36. [44]

    UHD: USRP hardware driver,

    ——, “UHD: USRP hardware driver,” GitHub repository, 2025, [On- line]. Available: https://github.com/EttusResearch/uhd. Accessed: Apr. 9, 2025

    2025

  37. [45]

    USRP hardware driver and USRP manual,

    ——, “USRP hardware driver and USRP manual,” https://files.ettus. com/manual/index.html, version 4.8.0.0. Accessed: Apr. 9, 2025

    2025

  38. [46]

    Time-varying system identification and channel equalization using wavelets and higher-order statistics,

    M. K. Tsatsanis, “Time-varying system identification and channel equalization using wavelets and higher-order statistics,” in Digital Signal Processing Systems: Implementation Techniques , ser. Control and Dynamic Systems, C. Leondes, Ed. Academic Press, 1995, vol. 68, pp. 333–394

    1995

  39. [47]

    Limits of reliable commu- nication with low probability of detection on AWGN channels,

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

    1921

Pith tools

Reviewed August 7, 2026 · model on record in the stance chip above.