{"id":"fef1c043-0e77-45d7-901c-ee21cf127a74","arxiv_id":"2506.20336","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":4.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":4,"one_line_summary":"A UAV-to-ground QKD model with grid-based photon capture finds that sub-10 cm beams and tens-of-micro-radian tracking are needed for Mbps raw key rates, with receiver FoV tuned to background and alignment conditions.","lead":"This paper develops an analytical model of UAV-to-ground quantum key distribution that includes beam spreading, transmitter and receiver misalignment, turbulence, and background light, and replaces the usual wide-beam approximation with a grid-based photon capture calculation.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Reported 'quantum key rate' in Fig. 4a is a raw click rate: Eq. (25) counts background-only slots as valid key bits and omits BB84 basis sifting, so the >2 Mbps secure-QKD claim is unsupported.","rationale":"The reader's weakest assumption exactly identifies the same load-bearing flaw: the raw-key model in Appendix D treats background-only detections as valid key bits and omits basis sifting. This is not a peripheral detail; it is the definition of the headline metric. The paper's stated contribution is a framework for evaluating QKD, and Eqs. (24)-(26) are the expressions used for all rate and QBER figures. Their misinterpretation means the abstract's claim of 'secure quantum key distribution ... Mbps-level key rates' is unsupported. I agree with the reader's conditional assessment: the paper's modeling insight about the wide-beam approximation is plausible and the grid-based approximation is validated against the exact integral and Monte Carlo, so the work is not without value. However, the performance results must be re-derived with a proper protocol-level key-rate model before the central claim can be accepted. Since the reader already assigned CONDITIONAL on exactly this basis, my stress test does not change the verdict.","tokens_in":19067,"tokens_out":3866,"duration_ms":37195,"concrete_test":"Recompute Figs. 4a and 4b with a standard BB84 sifted-key model: R_sifted = (1/2) * Rq * E[P(nq>=1, nb=0)] and QBER_sifted = E[P(nq>=1, nb=1)] / E[P(nq>=1, nb=0) + P(nq>=1, nb=1)], using the paper's own grid-based mu_p in Eq. (17) and the same simulation parameters. If R_sifted at sigma_thetae=50 urad and wz~5-10 cm falls significantly below 2 Mbps or QBER_sifted exceeds 1e-3, the abstract's secure-QKD rate claim is invalid and must be re-stated as a raw detection rate.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central performance claim depends on Proposition 4, Eq. (24)-(25), where Rkey = Rq * P(neff=1). Appendix D defines P(neff=1) as the probability that exactly one photon is detected, explicitly including State 2 (nb=1, nq=0, Eq. 44) as a raw key bit. In any real BB84-style protocol, a background-only detection carries no information about the transmitted bit and does not contribute to the sifted or secure key; it contributes to the error rate. Moreover, Eq. (24) contains no factor 1/2 for basis sifting. Consequently, the 'quantum key rate' reported in Fig. 4a and quoted in the abstract as 'Mbps-level key rates' is a detector click rate, not a QKD key rate. The QBER expression in Eq. (26) also counts only State 2 as erroneous, ignoring errors from basis mismatch, dark counts, and multi-photon WCP events. If the standard sifting factor is applied and background-only events are excluded, the headline rates fall by at least a factor of two and the QBER rises; the security claim would additionally require a decoy-state and finite-size analysis, neither of which is provided. The grid-based capture model itself is a valid contribution, but it does not rescue the performance metric as stated.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper presents an analytical framework for UAV-to-ground free-space quantum links, replacing the wide-beam Farid-Hranilovic pointing-error formula with a grid-based discretization of the exact Gaussian aperture-capture integral. It derives approximate closed-form expressions for single-photon detection probability, a quantity it calls the quantum key generation rate, and QBER, including transmitter-side FSM jitter, receiver angle-of-arrival misalignment, Gamma-Gamma turbulence, atmospheric attenuation, detector efficiency, FoV filtering, and background Poisson noise. The analytical results are validated against numerical integration of the exact integral and against Monte Carlo simulation over 10^6 time slots. The key design conclusions are that optimal beam waists lie below roughly 10 cm, that tracking precision must be at the microradian level, and that the receiver FoV must balance background rejection against misalignment tolerance; the paper reports Mbps-level key rates and QBER below 10^-3 for optimized parameters.","tokens_in":19421,"tokens_out":6835,"duration_ms":78134,"significance":"If accepted as a channel-capture model, the paper's grid-based approximation is a useful and credible contribution: the wide-beam approximation of Ref. [24] genuinely becomes inaccurate when the spot size is comparable to or smaller than the aperture, and the proposed discretization is shown to match the exact numerical integral and Monte Carlo with no fitted parameters. The paper also correctly identifies a real system-level trade-off between FoV, background noise, and receiver misalignment. However, the significance of the reported QKD performance numbers is currently compromised by the raw-click-rate definition of the key rate, by an algebraic inconsistency in the central formula, and by an unjustified linearization in the derivation of the detection probability. Until these are corrected, the abstract's 'secure QKD' and 'Mbps-level key rates' claims are not supported by the manuscript as written.","major_comments":[{"comment":"The quantity called Rkey is not a QKD sifted or secret key rate. Appendix D explicitly defines a raw key bit as any slot with exactly one detected photon 'regardless of its source,' and State 2 (nq=0, nb=1) contributes PS2=μ_b e^{-μ_b}[1-∫P(nq≥1|rd)f_rd(rd)drd]. Eq. (24) also contains no factor of 1/2 for BB84 basis sifting. A background-only detection carries no information about the transmitted bit and would enter a QKD protocol only as an error, not as a valid key bit; the >2 Mbps value in Fig. 4(a) and the abstract's 'Mbps-level key rates' are therefore detector click rates, not secure key rates. The QBER expression in Eq. (26) likewise counts only State 2 as erroneous and omits basis-mismatch errors and multi-photon weak-coherent-pulse events. The performance claims need to be re-derived with a protocol-level model, including sifting, error correction and privacy amplification, and ideally decoy-state and finite-size analysis, before they can be reported as secure QKD rates.","section":"Appendix D, Eqs. (24)-(25), Fig. 4"},{"comment":"There is an algebraic inconsistency in the central key-rate formula. From Appendix D, PS1=e^{-μ_b}S and PS3=(1/2)μ_b e^{-μ_b}S, where S denotes the integral over rd, so the coefficient multiplying S should be e^{-μ_b}(1+μ_b/2). Equation (25), however, contains e^{-μ_b}-(1/2)μ_b e^{-μ_b}, i.e., e^{-μ_b}(1-μ_b/2). The sign should be corrected, and the QBER expression in Eq. (26) should be checked against the corrected denominator. Although the discrepancy is numerically small when μ_b≪1, the formula as printed is not the sum of the three stated disjoint events.","section":"Eq. (25) versus Eqs. (43)-(45)"},{"comment":"The linearization e^{-z cpt μ_p} ≈ 1 - z cpt μ_p is applied after a change of variable in which z ranges over (0,∞), not only over small arguments. For z > 1/(cpt μ_p) the approximation becomes negative and does not approximate the exponential. The agreement with Monte Carlo in Fig. 3 is encouraging for the tested parameter range, but the derivation does not establish the conditions under which Proposition 3 remains accurate. The authors should either state and justify the low-transmissivity regime in which this approximation is valid or use the exact moment expression.","section":"Appendix C, Eqs. (22)-(23)"},{"comment":"The QBER definition considers only background-only slots as erroneous and assumes that State 1 and State 3 are error-free. In a polarization-encoded BB84 system, basis mismatch alone produces errors in a substantial fraction of sifted slots unless reconciliation is accounted for, and detector dark counts and multi-photon pulses also contribute. As written, Proposition 5 is not a QBER in the protocol sense and cannot be used to infer QKD security thresholds. This quantity should be either relabeled as a background-error indicator or extended to the full detection statistics.","section":"Eq. (26), Proposition 5"}],"minor_comments":[{"comment":"The symbol cpt is used in Propositions 2 and 3 before it is defined; it is first introduced in Appendix B. Define cpt = μ_t η_atm μ_d before Eq. (20).","section":"Section III, Propositions 2-3"},{"comment":"Table II sweeps wz directly, but Eq. (3) relates the received beam radius to the transmitter waist w0 and distance Lz. Clarify whether w0 is the optimized quantity and wz is computed from it, or whether the simulations independently set the spot radius at the receiver.","section":"Table II and Section IV"},{"comment":"The title and abstract promise 'adaptive FoV tuning strategies,' but Section IV only studies the FoV trade-off; there is no concrete adaptive algorithm or control law. Either add such a strategy or soften the claim.","section":"Title and Section IV"},{"comment":"The notation n_t for the number of transmitted photons and n_tot for the total number of detected photons is easy to confuse; use distinct symbols such as n_sig and n_tot.","section":"Appendix D"},{"comment":"Reference [18] is cited as a TechRxiv preprint; if a peer-reviewed version exists, it should be cited instead of or in addition to the preprint.","section":"References"}],"recommendation":"major_revision","confidential_remarks":null},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The short version: the paper's core modeling point is right, and the grid-based capture approximation is a legitimate, validated fix for the narrow-beam regime that UAV QKD actually needs. But the headline \"quantum key rate\" in Fig. 4 and the abstract is a raw detector click rate, not a QKD key rate, and that distinction matters.\n\nWhat's new: the observation that the Farid-Hranilovic wide-beam pointing-error model breaks down when wz is around 5–10 cm, and that this is exactly the regime you want for single-photon links, is correct and worth saying. The grid-based approximation in Proposition 1 is standard Riemann-sum quadrature, but it's implemented cleanly, validated against the exact 2D integral and Monte Carlo, and makes the subsequent analysis tractable. The design conclusions—sub-10 cm beam waists and micro-radian tracking—follow from the model and are plausible.\n\nSoft spots, in order of importance. First, the key rate metric. Proposition 4 and Appendix D define a raw key bit as any slot with exactly one detected photon, including State 2 where the only photon is a background photon. That is not a QKD key bit. BB84 discards half the bits in basis sifting, and a background-only detection carries no information about the transmitted bit—it is an error, not a key bit. So the greater-than-2 Mbps claim in Fig. 4a and the abstract's \"secure QKD\" framing are unsupported. The QBER expression is also incomplete: it only counts State 2 as erroneous and ignores basis mismatch, dark counts, and multi-photon WCP events. The authors do label Proposition 4 as \"raw key generation rate,\" but the abstract and conclusions do not keep that caveat. Second, the \"adaptive FoV tuning strategies\" promised in the abstract and contributions are not actually implemented; the paper sweeps FoV and shows the trade-off, which is fine, but the word \"adaptive\" oversells it. Third, minor: the Monte Carlo validation in Figs. 3–4 has no error bars, and the small-argument approximation in Appendix C (e^{-z cpt mu_p} ≈ 1 − z cpt mu_p) is not justified with numbers.\n\nThe math itself is sound as far as it goes—no fitted parameters, the grid model is checked against the exact integral rather than a fitted dataset, and the citation pattern is reasonable (self-citations are to the authors' own prior FSO work, which is relevant). This is not a case of a broken model; it is a case of a mismatched metric. The paper deserves a serious referee, but the performance claims need re-derivation with a real protocol model (sifting, decoy states, finite-size effects) before the secure-QKD conclusion can stand.\n\nRecommendation: send it to review, but instruct the referee to focus on the key rate definition and require a protocol-level correction.","headline":"Useful, correct fix to the wide-beam pointing-error model for narrow-beam UAV quantum links, but the headline key rates are raw click rates, not QKD key rates.","tokens_in":19910,"tokens_out":2248,"would_cite":false,"duration_ms":23355,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"UAV-to-ground quantum links need sub-10 cm beams and micro-radian tracking to exceed 2 Mbps key rates with QBER below 0.001; the paper's grid-based photon-capture model stays accurate where the classical wide-beam approximation fails.","keywords":["quantum key distribution","UAV-to-ground links","free-space optical communication","pointing errors","beam waist optimization","field-of-view control","quantum bit error rate","Gamma-Gamma turbulence"],"falsifier":"A direct check is to rerun the paper's own parameter set through a complete BB84-style protocol that discards roughly half the accepted detections for basis sifting, rejects background-only slots, and applies a decoy-state security analysis; if the resulting secret key rate no longer reaches the Mbps range or the QBER exceeds the protocol threshold, the central performance claim fails as stated. A complementary experimental test would measure the raw single-photon detection rate and QBER on a 1 km UAV-to-ground link with 50 micro-radian transmitter tracking error, beam waist 5–10 cm, and background radiance $10^{-6}$ W/m²/sr/nm, and compare the measured values with Propositions 4 and 5.","tokens_in":18810,"feed_emoji":"🔐","tokens_out":15337,"duration_ms":141219,"temperature":0.7,"pith_summary":"This paper tries to establish that the classical wide-beam approximation for pointing errors, which assumes the beam radius at the receiver is much larger than the aperture, becomes invalid in the quantum regime of UAV-to-ground links, where the beam waists that maximize photon collection are tightly focused (below 10 cm). The paper replaces that approximation with a grid-based discretization of the receiver aperture and derives analytical expressions for the average raw key generation rate and quantum bit error rate, folding in beam divergence, transmitter and receiver pointing errors, atmospheric attenuation, Gamma-Gamma turbulence, narrow field-of-view filtering, and background photon noise; a raw key bit is counted whenever exactly one photon is detected in a time slot, including slots where that photon is background light. On these modeling assumptions, optimized beam waists with 50 micro-radian transmitter tracking yield raw key rates above 2 Mbps and QBER below $10^{-3}$, while 2 milliradian tracking makes the link unusable with QBER above $10^{-1}$. The results matter because most existing UAV-QKD channel models inherit the wide-beam pointing-error formulation from classical free-space optics, and the paper shows that this inheritance fails exactly in the operating regime that makes airborne quantum key distribution feasible.","feed_headline":"Tight beams unlock Mbps raw-key rates for drone-to-ground QKD","feed_subtitle":"Grid-based model: sub-10 cm beams and micro-radian tracking hit 2 Mbps with QBER below 0.001.","key_machinery":"The load-bearing object is the grid-based photon capture approximation (Proposition 1). For a Gaussian single-photon probability density and a circular receiver aperture of radius $r_a$, the exact capture probability is a two-dimensional integral over the aperture; the grid scheme splits the aperture into $N_g$ segments of width $\\Delta x=2r_a/N_g$, evaluates the Gaussian at each segment center $x_i$, and folds the segment's vertical extent into the coefficient $c_i = (2\\Delta x/\\sqrt{2\\pi} w_z)\\,\\mathrm{erf}(\\sqrt{2/w_z^2}\\sqrt{r_a^2-x_i^2})$. This turns the capture probability into a sum of one-dimensional Gaussian terms that remain accurate when the beam waist $w_z$ is comparable to or smaller than the aperture, which is precisely the regime where the classical wide-beam formula $\\mu_p(r_d)\\approx (2r_a^2/w_z^2)\\exp(-2\\|r_d\\|^2/w_z^2)$ fails. The capture probability then enters the end-to-end mean detected photon count $\\mu_q = \\mu_t \\eta_{atm} \\mu_d \\mu_p(r_d) \\eta_{turb} \\mu_{FoV}$, with $\\mu_{FoV}$ a Bernoulli visibility factor determined by the receiver's angular acceptance cone. Averaging over the Rayleigh-distributed beam displacement $r_d$ and the Gamma-Gamma turbulence $\\eta_{turb}$ reduces the key rate and QBER to single one-dimensional integrals, which is what makes the framework tractable enough for system-level optimization.","core_discovery":"The central claim, stated in the paper's own terms, is that 'the wide-beam approximation in [24] becomes invalid' for UAV-to-ground quantum links, and that replacing that approximation is required to design links that actually work. The paper derives a grid-based approximation for the photon capture probability $\\mu_p(r_d)$ that divides the receiver aperture into $N_g$ equal strips and represents each strip's contribution as a weighted Gaussian, so that $\\mu_p(r_d) \\approx \\sum_{i=1}^{N_g} c_i \\exp(-2(x_i-r_d)^2/w_z^2)$, with coefficients $c_i$ containing an error function; with $N_g=10$ it tracks exact numerical integration. This capture probability feeds a Poisson model for detected photons per quantum slot, and the paper's Propositions 2–5 produce one-dimensional integral expressions for the probability of detecting at least one photon, the raw key rate $R_{key}=R_q P(n_{\\mathrm{eff}}=1)$, and the QBER. The headline quantitative results are that for 50 micro-radian transmitter tracking error the raw key rate surpasses 2 Mbps and QBER drops below $10^{-3}$, while for 2 milliradian tracking error the key rate falls below 100 kbps and QBER exceeds $10^{-1}$, and that the receiver field of view must be tuned as a trade-off between signal capture and background rejection.","pith_inferences":["A full protocol-level analysis with basis sifting and decoy states would likely reduce the quoted Mbps figures by a factor of two or more, and possibly much more, because the paper's raw-key definition counts background-only slots as key bits; this is an inference from the paper's Appendix D, not its stated conclusion.","The grid-based aperture model could be combined with an online estimator of background radiance and pointing-error variance to close the loop on FoV control; the paper proposes adaptive FoV tuning but does not specify a control algorithm.","For very strong turbulence or non-Gaussian beam distortions, the Gamma-Gamma and Gaussian-beam assumptions may need re-examination, so the quantitative predictions are safest in the moderate-turbulence regime the paper assumes.","A natural next step is to extend the framework to entangled-photon UAV links, where two correlated channels must be modeled jointly and background rejection matters even more; the paper mentions entanglement as future work."],"forward_implications":["Existing UAV-QKD analyses that use the wide-beam pointing-error model of [24] should be re-examined, because the approximation is inaccurate for the sub-10 cm beam waists the paper identifies as optimal.","Link designers should treat receiver field of view as an adaptive parameter: wider FoV raises key rate under low background light, while narrower FoV is needed to keep QBER acceptable under daylight.","The analytical framework enables fast joint optimization of beam waist, tracking precision, photon emission rate, and FoV without full Monte Carlo simulation.","Milliradian-level tracking, which is acceptable for classical FSO links, makes UAV-to-ground QKD unusable in this model; micro-radian tracking is necessary.","The same modeling approach applies with minor modifications to ground-to-UAV and terrestrial short-range free-space QKD, as the paper states in its conclusion."],"supporting_citations":[{"why":"Supplies the classical wide-beam pointing-error approximation that the paper argues becomes invalid under tightly focused beams.","marker":"[24]"},{"why":"Provides the Gamma-Gamma probability density for turbulence-induced transmittance used in the photon-count distribution and key-rate/QBER integrals.","marker":"[32]"},{"why":"Gives the Poisson photon-number statistics of the weak coherent pulse source used to model transmitted signal photons.","marker":"[27]"},{"why":"Supplies the Poisson model for background photon counts per time slot that underlies the noise terms in the key rate and QBER expressions.","marker":"[33]"},{"why":"Provides the integral identity used to evaluate the moment integral over Gamma-Gamma transmittance when deriving the closed-form approximations.","marker":"[36]"},{"why":"Gives the Gaussian beam propagation formula and Beer-Lambert atmospheric transmittance used throughout the channel model.","marker":"[29]"},{"why":"Supplies the bivariate Gaussian model for receiver angle-of-arrival misalignment that defines the FoV visibility factor.","marker":"[34]"},{"why":"Models the transmitter fast-steering-mirror residual pointing error as zero-mean Gaussian angular jitter, converted into beam displacement.","marker":"[30]"}],"fun_headline_variants":["Grid-based model shows tight beams enable Mbps UAV QKD","Sub-10 cm beams and micro-radian tracking hit Mbps QKD rates","Adaptive FoV tuning balances noise and misalignment in drone QKD","UAV QKD needs tight beams: grid model predicts 2 Mbps key rates","Grid approximation replaces wide-beam for drone-to-ground QKD"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that every time slot in which exactly one photon is detected counts as a valid raw key bit, even when that photon is stray background light, and no fraction of detections is set aside for the sender-receiver basis comparison that real quantum protocols perform; if those real-protocol steps are included, the reported key rates would fall.","fun_headline_variants_meta":{"raw":{"variants":["Grid-based model shows tight beams enable Mbps UAV QKD","Sub-10 cm beams and micro-radian tracking hit Mbps QKD rates","Adaptive FoV tuning balances noise and misalignment in drone QKD","UAV QKD needs tight beams: grid model predicts 2 Mbps key rates","Grid approximation replaces wide-beam for drone-to-ground QKD"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000741,"raw_usage":{"total_tokens":3363,"prompt_tokens":1058,"completion_tokens":2305,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":674,"completion_tokens_details":{"reasoning_tokens":2212}},"tokens_in":674,"tokens_out":2305,"duration_ms":16597,"temperature":1.0,"reasoning_tokens":2212,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-06T22:52:55.217869+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"A direct check is to rerun the paper's own parameter set through a complete BB84-style protocol that discards roughly half the accepted detections for basis sifting, rejects background-only slots, and applies a decoy-state security analysis; if the resulting secret key rate no longer reaches the Mbps range or the QBER exceeds the protocol threshold, the central performance claim fails as stated. A complementary experimental test would measure the raw single-photon detection rate and QBER on a 1 km UAV-to-ground link with 50 micro-radian transmitter tracking error, beam waist 5–10 cm, and background radiance $10^{-6}$ W/m²/sr/nm, and compare the measured values with Propositions 4 and 5.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the Gamma-Gamma probability density for turbulence-induced transmittance used in the photon-count distribution and key-rate/QBER integrals."},{"cited_title":"Security bounds for decoy-state quantum key distribution with arbitrary photon-number statistics,","cited_arxiv_id":null,"evidence_quote":"Gives the Poisson photon-number statistics of the weak coherent pulse source used to model transmitted signal photons."},{"cited_title":"Photon-counting statistics-based support vec tor machine with multi-mode photon illumination for quantum imaging,","cited_arxiv_id":null,"evidence_quote":"Supplies the Poisson model for background photon counts per time slot that underlies the noise terms in the key rate and QBER expressions."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the integral identity used to evaluate the moment integral over Gamma-Gamma transmittance when deriving the closed-form approximations."},{"cited_title":"CRC press, 2019","cited_arxiv_id":null,"evidence_quote":"Gives the Gaussian beam propagation formula and Beer-Lambert atmospheric transmittance used throughout the channel model."},{"cited_title":"A Novel MRR-UA V -Based Relay W ith Optical Network Coding: A Comparative Study With Optical IR S and Conventional UA V Relaying,","cited_arxiv_id":null,"evidence_quote":"Models the transmitter fast-steering-mirror residual pointing error as zero-mean Gaussian angular jitter, converted into beam displacement."}],"review_version":1}