{"id":"155b286a-a254-4ed5-861d-d7b191e17695","arxiv_id":"2506.07831","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":4,"one_line_summary":"A two-stage protocol, GPS pulse alignment followed by entanglement-based temporal correction, achieves 24 ± 12 ps RMS clock drift in a lab emulation of a dynamic free-space drone channel.","lead":"The paper proposes a two-stage clock synchronization scheme for drone-based quantum key distribution: coarse timing from GPS pulses, then fine correction using the arrival times of entangled photon pairs. A tabletop experiment with a moving optical delay reports 24 ± 12 ps clock drift over one hour, suggesting drones can maintain the precise timing that entanglement-based QKD requires without carrying heavy atomic clocks.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The 24±12 ps headline is measured on the same corrected coincidence peaks the protocol itself adjusts, so it validates loop residual, not synchronization accuracy; an independent reference test is needed.","rationale":"The reader's weakest_assumption is exactly the load-bearing point: the headline precision is inferred from the corrected coincidence peak, the same signal the algorithm minimizes. I agree with that diagnosis. I also note the Eq. 13/Appendix D inconsistency independently weakens the paper's statistical-support argument. A single external-clock experiment would settle whether 24±12 ps is real. Since the concern is not fatal—the tabletop setup and data are plausible—the appropriate disposition remains CONDITIONAL, so I recommend no change to the reader's verdict.","tokens_in":14890,"tokens_out":5371,"duration_ms":67533,"concrete_test":"Re-run the 1-hour 7.5 dB loss experiment with both TTMs disciplined by a common, independently calibrated 10 MHz reference (or a rubidium clock) and compare the corrected coincidence-peak offset time series against that reference; separately, compute the same 24±12 ps statistic from raw (uncorrected) coincidence peaks over the same dataset. If the residual against the independent reference exceeds the claimed 24 ps, or if the no-correction baseline already shows comparable stability, the headline claim is not supported by the present data.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim (abstract, §V) is that the two-stage protocol achieves 24±12 ps RMS clock drift under 300 ps/s delay dynamics. The only drift metric reported, Fig. 6(c), is the temporal offset of the corrected coincidence peak. This is the same observable that TSEC (Eqs. 8–11) uses to re-estimate and correct the timestamps; the residual therefore measures the closed-loop error of an estimator applied to its own training data, not agreement with any independent time base. §IV states TTM jitter is 45 ps RMS and GNSS PPS jitter is 10–20 ns, so without an external reference it is impossible to distinguish genuine picosecond clock agreement from a smoothed, low-bandwidth residual. The emulated dynamics are also a constant 300 ps/s ramp; TSEC's per-sub-block linear interpolation can track a constant velocity exactly, so the experiment does not exercise the acceleration term that Eq. 12 is meant to bound. The internal photon-pair bound is additionally inconsistent: Eq. 13 gives Ns≥0.5 under practical conditions, while Appendix D (Eq. D6) gives N≥15.36 for the stated typical parameters; the 219 cps per channel pair with S=20 yields ~11 pairs per sub-block, sitting between these conflicting requirements. These issues do not prove the protocol is wrong, but they make the headline precision unanchored.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript proposes a two-stage clock-synchronization protocol for entanglement-based QKD on drones: coarse alignment using GNSS PPS time tags, followed by temporal-sliced entanglement correction (TSEC) that re-estimates the coincidence-peak offset in short sub-blocks and linearly interpolates the clock correction between them. The authors derive error bounds for PPS-TA and TSEC, present simulations of sub-block partitioning, and report a tabletop experiment in which an optical delay line emulates a 300 ps/s relative delay ramp and a variable attenuator emulates channel loss. Over a 1-hour run they report a residual clock drift of 24 ± 12 ps RMS at coincidence rates down to 219 cps per channel pair, together with a dual-pointer correlation algorithm that runs in milliseconds.","tokens_in":15067,"tokens_out":7590,"duration_ms":79277,"significance":"The practical goal—picosecond-class synchronization on SWaP-constrained drone platforms using only GNSS PPS and a low-stability crystal oscillator—is well motivated, and the tabletop apparatus is a reasonable physical emulation of the loss and delay dynamics. The paper contains a genuine experimental implementation with a polarization-entangled source, optical delay line, variable attenuator, SNSPDs, and a 1-hour stability run; the O(n) dual-pointer correlation algorithm with measured 7.49 ± 0.74 ms processing time is a useful engineering contribution. If the synchronization precision were validated against an independent time base, the protocol would be a meaningful step toward mobile entanglement-based QKD. However, the headline 24 ± 12 ps figure is currently not anchored: it is measured on the same corrected coincidence peaks that the protocol itself adjusts, and the supporting error analysis contains inconsistencies that need to be resolved before the central claim can be accepted.","major_comments":[{"comment":"The reported 24 ± 12 ps RMS 'clock drift' is derived from the temporal offset of the corrected coincidence peak, which is the same observable that the TSEC stage (Eqs. 8–11) uses to re-estimate and correct the timestamps. The residual therefore characterizes the closed-loop error of an estimator applied to its own training observable, not agreement with an independent time base. Given that the TTM has 45 ps RMS jitter and the GNSS PPS has 10–20 ns jitter, no measurement in the paper rules out the possibility that the 24 ps figure reflects a smoothed, low-bandwidth residual rather than genuine picosecond clock agreement. Please add an independent validation, for example a loop-back measurement against a common calibrated reference clock or a comparison of Alice's and Bob's time tags against a disciplined external timebase, and report the raw (unfiltered) residual distribution as well as the MAD-filtered one.","section":"Section IV, Fig. 6(c)"},{"comment":"The error-propagation derivation for PPS-TA is not dimensionally consistent. In Eq. (A3), the first term is dimensionless while the second term has units of time squared if t_Sync is treated as a normalized fraction; in Eq. (A4) the second term contains an explicit factor n that contradicts the statement that 'all errors scale as 1/n'. The text also states that tSync_i ∈ [0,1] immediately after defining tSync in Eq. (3) as a quantity that ranges over the alignment window. Please rederive this propagation with consistent definitions and verify the 1/n scaling against a Monte Carlo simulation of the PPS resampling.","section":"Appendix A, Eqs. (A3)–(A4)"},{"comment":"The minimum photon-pair requirements are inconsistent. Equation (13) gives Ns ≥ 0.5 under the stated assumptions, whereas Appendix D, Eq. (D6), with the paper's own typical parameters σ_TSEC = 100 ps and γ = 50 ps, gives N ≥ 15.36. The 1-hour experiment operates at 219 cps per channel pair; with S = 20 and a 1 s block, this is about 11 pairs per sub-block, which lies between these two conflicting bounds. The paper should state which bound is the operative one for the claimed 24 ps precision and should reconcile the two derivations, since the statistical significance of the sub-block mean is load-bearing for the TSEC claim.","section":"Section III.B.2, Eq. (13); Appendix D, Eq. (D6)"},{"comment":"The dynamic test uses an optical delay line that produces a constant 300 ps/s relative delay ramp, i.e., a constant velocity with zero acceleration. TSEC's per-sub-block linear interpolation can track a constant ramp exactly, so the experiment does not exercise the acceleration term that Eq. (12) is designed to bound, and the claim of robustness under 'distance dynamics' should be limited to constant-velocity motion unless an accelerated scenario is also tested. The conversion in Table I of γ = 0.3 ps/ms² as 'equivalent to 0.1 m/s channel distance variation' also mixes acceleration and velocity units and should be corrected.","section":"Section IV and Eq. (12)"}],"minor_comments":[{"comment":"Equation (2) writes pA_m = pG_m + σ²_p and qB_m = qG_m + σ²_q; this should be pA_m = pG_m + ε_p with ε_p ~ N(0, σ²_PPS), since a variance is not added to a time.","section":"Eq. (2)"},{"comment":"The expression for σ²_TS is dimensionally inconsistent: |∂(ΔtB_j)/∂tB_j| σ_tB has units of time, while (σ_tB_i/Δτ_s)² is dimensionless. Please correct the units and the definition.","section":"Eq. (10)"},{"comment":"The phrase 'The correction coincidence count rate' should read 'the corrected coincidence count rate', and 'Fig. 6(b) characteristic the clock drift' should read 'characterizes the clock drift'.","section":"Section IV"},{"comment":"The sentence 'it simultaneously decreases the number of entangled of entangled photon pairs per sub-block' contains a duplicated phrase and should be edited.","section":"Section III.B.2"},{"comment":"The caption should explicitly define how the clock drift is estimated from the coincidence-peak offset and state whether the plotted values are raw or MAD-filtered residuals, since 2.83% of samples were removed before reporting the 24 ± 12 ps value.","section":"Fig. 6(c) caption"},{"comment":"The phrase 'without requiring precision reference clock' should be clarified to mean that no precision clock is required at the drone nodes; the laboratory validation of the claimed precision would still benefit from an independent reference, as noted in the major comments.","section":"Abstract and Section V"}],"recommendation":"major_revision","confidential_remarks":"The manuscript is a reasonable experimental contribution, but the headline precision claim is not yet supported. I recommend sending it back with a request for an independent timing-reference measurement and a corrected error analysis. The inconsistency between Eq. (13) and Appendix D should be resolved before the statistical claims are taken as established. No concerns about novelty or attribution beyond scope were identified."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Quick take: this is a solid engineering demonstration with a real entangled source and a sensible two-stage scheme, but the headline number is not yet anchored. The protocol is not conceptually new—GNSS coarse alignment plus entanglement-based correction appears in the cited literature—but applying it to a drone SWaP context with dynamic delay emulation and a fast correlation algorithm is genuinely useful.\n\nWhat's good: they built the hardware (0.3 kg board, drone-mountable), ran a 1-hour stability test, varied loss up to 7.5 dB, and got stable coincidence counts at 219 cps per channel pair. The dual-pointer O(n) correlation is a practical contribution. The literature table is honest about prior GNSS+EC work.\n\nThe soft spots are significant though. The 24±12 ps RMS 'clock drift' in Fig. 6(c) is derived from the offset of the coincidence peaks that TSEC itself re-estimates and corrects. That's the loop's residual error, not agreement with an independent time base. The paper never compares against a reference clock, and with TTM jitter at 45 ps it's hard to believe 24 ps absolute unless the correction truly cancels common-mode noise. The emulated dynamics are a constant 300 ps/s ramp, which a linear per-sub-block fit tracks exactly; the protocol's acceleration term (Eq. 12) is never exercised. Also, the minimum pair bounds disagree: Eq. 13 gives N_s ≥ 0.5, Appendix D gives N ≥ 15.36 under typical parameters, and the 219 cps with S=20 yields ~11 pairs per sub-block, sitting between them. The outlier removal (2.83% of points dropped) needs justification and a no-outlier-removal version.\n\nNone of this proves the method wrong—the key and QBER data are self-consistent—but the central precision claim overreaches relative to what was measured. A revision with an external reference (even a second TTM on a common clock), a no-correction baseline, and reconciled resource bounds would make it convincing.\n\nWho is this for? People working on mobile quantum networks and SWaP-constrained QKD terminals will find the engineering useful. I'd send it to a serious referee—it deserves engagement—but I would not cite the 24 ps figure until it's independently confirmed.","headline":"A competent tabletop demo of a plausible two-stage sync protocol, but the 24 ps headline measures the loop's residual, not synchronization accuracy against an independent clock.","tokens_in":15740,"tokens_out":1819,"would_cite":false,"duration_ms":21200,"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":"GNSS pulses plus entanglement correlations hold drone QKD clocks to 24 ± 12 ps RMS.","keywords":["clock synchronization","entanglement-based QKD","drone quantum networks","GNSS PPS timing","temporal-sliced entanglement correction","coincidence measurement","timing jitter","SWaP constraints"],"falsifier":"Run the protocol at both ends of a moving link while each node also logs time against an independent trusted clock, such as a rubidium reference, and make the link distance accelerate and jerk instead of moving at constant speed; if the external clock disagrees with the protocol's corrected timestamps by much more than 24 ps, the claimed precision is an artifact of self-referencing.","tokens_in":14568,"feed_emoji":"⏱️","tokens_out":13876,"duration_ms":150363,"temperature":0.7,"pith_summary":"This paper claims that two low-precision, drone-compatible timing ingredients—a GNSS pulse-per-second signal and the arrival-time correlations of entangled photon pairs—can replace a dedicated precision clock in mobile entanglement-based quantum key distribution. It proposes a two-stage protocol: first align timestamps coarsely to the GNSS pulse, then split the coincidence data into short temporal blocks and use each block's coincidence-peak offset to correct residual drift. In a tabletop free-space channel with the optical delay swept at 300 ps/s to mimic motion, the protocol holds clock drift at $24\\pm12$ ps RMS over one hour, at coincidence rates as low as 219 counts per second per channel pair and channel losses up to 7.5 dB. If true, drone platforms can keep the picosecond-level timing that entanglement measurements need without rubidium clocks or dedicated synchronization pulses, which matters because timing jitter inflates accidental coincidences and cuts secure key rates.","feed_headline":"Two-stage timing fix holds drone QKD clocks to 24 ps","feed_subtitle":"A low-grade GNSS receiver plus entanglement-based correction keeps picosecond sync while the link moves.","key_machinery":"The central object is the temporal offset of the entangled-photon coincidence peak: because the two photons of a pair are detected at different nodes at times whose difference equals the propagation delay plus the clock offset, the position of the intensity-correlation peak continuously encodes the relative clock error. The machinery has two stages: PPS-TA uses the GNSS pulse sequence to align timestamps coarsely at nanosecond accuracy, and TSEC slices the coincidence histogram into temporal sub-blocks, computes the mean offset $\\bar\\mu_s$ for each slice, and linearly interpolates a correction to every timestamp, driving residual jitter down to the level set by photon statistics and timestamp-transfer error. The argument is carried by the error-propagation equations, the linearization criterion $S_{\\min}$ set by drift acceleration, and the minimum-pair bound $N_s$ from Cramér–Rao analysis.","core_discovery":"On its own terms, the discovery is that residual timing error left over after a coarse GNSS alignment can be pushed to tens of picoseconds by using the entangled photons themselves as a continuously updated clock reference, even while the channel length changes. The protocol, coarse pulse-per-second alignment (PPS-TA) followed by temporal-sliced entanglement correction (TSEC), treats a one-second block of coincidence data as a sequence of sub-blocks in which the clock drift is approximately linear; the mean offset $\\bar\\mu_s$ of each sub-block is estimated from the intensity-correlation peak, and a linear interpolation assigns a corrected time to every detected photon. The central experimental result is a measured $24\\pm12$ ps RMS clock drift over one hour against a programmed 300 ps/s relative delay, with no observable increase in timing jitter as loss rose to 7.5 dB and a sustained 259 bps secure key rate at a 5.63% quantum bit error rate. The paper also derives resource bounds—a minimum sub-block count set by drift acceleration and a minimum entangled-pair count per block from a Cramér–Rao analysis—and reports a dual-pointer correlation algorithm that processes both timestamp streams in $7.49\\pm0.74$ ms on a low-power mini-PC, making the correction loop fast enough for real-time use.","pith_inferences":["The 24 ps figure is read from the same coincidence-peak offset the protocol corrects, so it measures residual noise of the feedback loop rather than agreement with an independent clock; anchoring both nodes to an external trusted time reference would separate absolute clock error from photon-statistics noise.","The experiment emulates motion as a constant 0.1 m/s delay ramp, so a field trial with acceleration, attitude jitter, and GNSS multipath would directly stress the linear-drift assumption that sub-blocking relies on.","Because the coincidence peak is available in any entanglement-distribution link, the same two-stage idea should transfer to other distributed quantum tasks needing remote time alignment, such as Bell-state measurements for quantum repeaters or entanglement swapping between mobile nodes.","The $S_{\\min}$ bound suggests an adaptive variant that shortens temporal sub-blocks during fast maneuvers and lengthens them in stable cruise, balancing linearization error against photon statistics per block."],"forward_implications":["Drone-to-drone entanglement QKD can operate with a 20 cm by 20 cm, 0.3 kg timing board built around a nanosecond GNSS receiver and a $\\pm1.0$ ppm/year oscillator, with no precision reference clock.","At 7.5 dB channel loss and 219 cps coincidence rate per channel pair, the protocol still yielded a 259 bps secure key rate at 5.63% quantum bit error rate, so timing is not the limiting factor at practical loss levels.","Real-time correction is compatible with constrained onboard computing: the dual-pointer algorithm processed $144\\times10^3$ and $133\\times10^3$ detection events in $7.49\\pm0.74$ ms on a low-power mini-PC.","The reported $24\\pm12$ ps precision lands in the same range as systems built on rubidium clocks or dedicated synchronization pulses, but without those hardware requirements."],"supporting_citations":[{"why":"Supplies the 13 ps/s drift figure for a GNSS-disciplined rubidium clock that motivates comparing this low-stability-oscillator result against precision references.","marker":"[3]"},{"why":"Demonstrates GNSS plus entanglement correction with sub-1.6 ns coincidence windows in free space, the architecture the PPS-TA stage extends to drone motion.","marker":"[7]"},{"why":"Continues the same GNSS-and-entanglement-correction approach at 270 m free-space links, a baseline for coincidence-window refinement.","marker":"[8]"},{"why":"Shows entanglement-assisted correction with a rubidium clock reaching below 12 ps over 50 km, the precision benchmark this protocol approaches without a rubidium clock.","marker":"[15]"},{"why":"Establishes that the temporal shift of the coincidence peak between two distant nodes encodes their relative clock offset, the principle TSEC reads.","marker":"[22]"},{"why":"Demonstrates remote clock synchronization from the arrival-time correlation of entangled photon pairs, directly supporting the TSEC measurement principle.","marker":"[23]"},{"why":"Provides the low-coincidence-rate regime in which the paper claims stable operation at 219 cps per channel pair.","marker":"[27]"},{"why":"Shows entanglement-corrected synchronization below 68 ps on a fixed free-space link, the static result this protocol extends to dynamically varying delay.","marker":"[28]"}],"fun_headline_variants":["Entanglement timing correction syncs drone QKD to 24 ps","GNSS plus entangled photons: 24 ps sync for drone QKD","Drone QKD hits 24 ps clock sync with mixed GNSS-photon timing","24 ps RMS sync on moving drones using entanglement"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that the corrected photon-pair coincidence peak is a faithful readout of true clock disagreement, and that the laboratory's steady 300 ps/s delay ramp represents real drone motion; if either fails, the 24 ps figure may not transfer to flight.","fun_headline_variants_meta":{"raw":{"variants":["Entanglement timing correction syncs drone QKD to 24 ps","GNSS plus entangled photons: 24 ps sync for drone QKD","Drone QKD hits 24 ps clock sync with mixed GNSS-photon timing","24 ps RMS sync on moving drones using entanglement"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000305,"raw_usage":{"total_tokens":1758,"prompt_tokens":959,"completion_tokens":799,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":575,"completion_tokens_details":{"reasoning_tokens":723}},"tokens_in":575,"tokens_out":799,"duration_ms":9664,"temperature":1.0,"reasoning_tokens":723,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-07T05:24:52.160505+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Run the protocol at both ends of a moving link while each node also logs time against an independent trusted clock, such as a rubidium reference, and make the link distance accelerate and jerk instead of moving at constant speed; if the external clock disagrees with the protocol's corrected timestamps by much more than 24 ps, the claimed precision is an artifact of self-referencing.","supporting_citations":[{"cited_title":"Smin can be determined based on σTSEC within each integration period Ts","cited_arxiv_id":null,"evidence_quote":"Supplies the 13 ps/s drift figure for a GNSS-disciplined rubidium clock that motivates comparing this low-stability-oscillator result against precision references."},{"cited_title":"The QBER induced by clock uncertainty is γT/2","cited_arxiv_id":null,"evidence_quote":"Demonstrates GNSS plus entanglement correction with sub-1.6 ns coincidence windows in free space, the architecture the PPS-TA stage extends to drone motion."},{"cited_title":"Wengerowsky, S","cited_arxiv_id":null,"evidence_quote":"Continues the same GNSS-and-entanglement-correction approach at 270 m free-space links, a baseline for coincidence-window refinement."},{"cited_title":"Basso Basset, M","cited_arxiv_id":null,"evidence_quote":"Shows entanglement-assisted correction with a rubidium clock reaching below 12 ps over 50 km, the precision benchmark this protocol approaches without a rubidium clock."},{"cited_title":"Entanglement-based clock syntonization for quantum key distribution networks. Demonstration over a 50 km-long link","cited_arxiv_id":"2501.16796","evidence_quote":"Establishes that the temporal shift of the coincidence peak between two distant nodes encodes their relative clock offset, the principle TSEC reads."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Demonstrates remote clock synchronization from the arrival-time correlation of entangled photon pairs, directly supporting the TSEC measurement principle."},{"cited_title":"Gerrits, I","cited_arxiv_id":null,"evidence_quote":"Provides the low-coincidence-rate regime in which the paper claims stable operation at 219 cps per channel pair."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Shows entanglement-corrected synchronization below 68 ps on a fixed free-space link, the static result this protocol extends to dynamically varying delay."}],"review_version":1}