{"id":"c5adc39f-0466-47f5-b002-698adfe3a5af","arxiv_id":"2607.11620","paper_version":1,"verdict":"CONDITIONAL","confidence":"HIGH","novelty_score":6.0,"correctness_risk":"low","formal_verification":"none","parameter_count":4,"one_line_summary":"SIC lets a narrow-beam LEO uplink serve three effective UEs per cell with finite SIR variance and only a 0.2 bit/s/Hz throughput penalty versus single-user service.","lead":"This paper derives SIR order statistics for the strongest users at a LEO satellite beam under successive interference cancellation, using stochastic geometry and a mixture-exponential shadowing model. It shows that serving three effective users per beam with SIC trades a modest throughput drop for finite SIR variance and better fairness.","discovery_kind":"new_application","skeptic_critique":{"model":"grok-4.5","headline":"No significant objection identified beyond the reader's already-flagged moment-matching approximation.","rationale":"The reader's weakest-assumption diagnosis is precisely the load-bearing modeling step. Because the paper already supplies spherical Monte-Carlo validation that employs the un-approximated Gaussian mixture and shows the analytic curves lying inside the simulated bands, the approximation error is already constrained for the headline numbers. No deeper algebraic inconsistency or hidden divergence appears in the Poisson-Dirichlet derivation or the SIC indicator conditions once the gamma characterization is granted. Consequently the conditional-acceptance verdict stands; the only useful next check is a direct numerical quantification of the residual error of the moment-matched Laplace transform, which the concrete_test isolates.","tokens_in":9585,"tokens_out":528,"duration_ms":5148,"concrete_test":"Recompute the three-user SIC coverage curves of Fig. 4 and the throughput integral (26) by replacing the mixture-exponential Laplace transform with a numerical Laplace transform obtained from the true two-component log-normal mixture (Eq. 2) via Monte-Carlo or quadrature; if any of the three coverage probabilities at θ=−7 dB shifts by more than 0.05 or T3 moves outside [0.65,0.75] bit/s/Hz, the quantitative claim weakens.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim (SIC with κυ=n=3 yields ~900% third-UE coverage gain at θ=−7 dB, total throughput only 0.9→0.7 bit/s/Hz, and finite SIR variance) rests on the mixture-exponential approximation (Eqs. 3–5) that matches the first two moments of the 3GPP urban Gaussian-mixture shadowing. All subsequent objects—Laplace transform of I becoming gamma (Eq. 12), Poisson-Dirichlet PD(0,κυ̃) densities (Prop. 2), joint order-statistic PDFs (Eq. 18), and the SIC coverage integrals (Prop. 3 and Eq. 26)—are exact only under that approximation. The paper already validates the final coverage curves against spherical Monte-Carlo that uses the true Gaussian mixture (Figs. 3–4, shaded bands for ϵ∈[30°,90°]), so the approximation error is empirically bounded for the reported operating point. Perfect SIC and the planar narrow-beam idealization are secondary modeling choices already noted by the reader; they do not introduce an independent, load-bearing inconsistency with the strongest claim.","agreement_with_reader":"agree"},"referee_report":{"model":"grok-4.5","summary":"The paper studies SIR order statistics and successive interference cancellation (SIC) for the uplink of a typical narrow-beam LEO satellite base station. UEs form a homogeneous PPP; the receive beam is Gaussian. Urban shadowing is taken from a two-tier Gaussian mixture (3GPP parameters) and approximated by a mixture-exponential distribution whose single free parameter υ is fixed by matching the first two moments. Under this model the total received power is gamma-distributed, the STIR process is Poisson–Dirichlet PD(0,κυ̃), and closed-form factorial-moment densities together with the joint PDF of the ordered STIRs are obtained. Coverage probabilities for the three strongest UEs with and without perfect SIC, as well as the aggregate bandwidth-normalized throughput, are then evaluated. Analytic curves for κυ = n = 3 are shown to lie inside Monte-Carlo envelopes generated on a spherical Earth with the original Gaussian-mixture shadowing; the results indicate that SIC raises third-UE coverage at θ = −7 dB by roughly 900 % while total throughput falls only from ≈0.9 to ≈0.7 bit/s/Hz and keeps SIR variance finite.","tokens_in":9877,"tokens_out":1349,"duration_ms":22895,"significance":"If the quantitative claims hold, the work supplies the first stochastic-geometry treatment of SIC order statistics that is specific to narrow-beam LEO uplinks and that correctly identifies the underlying gamma-process / PD(0,·) structure (distinct from the terrestrial PD(α,0) case). The explicit Laplace transform of interference, the factorial-moment densities, and the SIC coverage integral are immediately usable design formulae. The concrete trade-off—finite SIR variance and multi-UE service at a modest aggregate-throughput cost—is a falsifiable prediction that is already checked against spherical Monte-Carlo with 3GPP shadowing parameters. The derivation chain from the PPP mapping theorem through moment-matched fading to the gamma/PD objects is rigorous and transparent.","major_comments":[{"comment":"§II.B.2, Eqs. (4)–(5) and subsequent use in (12), Prop. 2 and Prop. 3: all analytic objects (Laplace transform of I, PD densities, SIC coverage integrals) are exact only under the mixture-exponential approximation that matches the first two moments of the 3GPP Gaussian mixture. Although Figs. 3–4 show that the final coverage curves fall inside the spherical Monte-Carlo bands that employ the true mixture, a direct comparison of the Laplace transform (or of the empirical CCDF of total interference I) under both fading models would quantify residual approximation error beyond mean/variance matching and would make the central numerical claims more robust.","section":"§II.B.2 / Eqs. (4)–(5)"},{"comment":"§III.C, Prop. 3 and Eq. (26): the reported 900 % coverage gain and the 0.7 bit/s/Hz throughput rest on the assumption of perfect successive cancellation. The paper notes that imperfect cancellation is “straightforward” to include, yet no residual-interference factor or sensitivity plot is provided. Because the fairness/throughput trade-off is the main design takeaway, even a simple one-parameter residual model (or a lower bound) would strengthen the claim that the multi-UE operating point remains attractive under realistic SIC.","section":"§III.C / Prop. 3"}],"minor_comments":[{"comment":"Table I lists σ^{2}_LoS = 42 [dB] and σ^{2}_NLoS = 62 [dB], while the text correctly states σ_LoS = 4 dB and σ_NLoS = 6 dB. The table entries are almost certainly typesetting artefacts for 4^{2} and 6^{2}; they should be corrected to avoid confusion.","section":"Table I"},{"comment":"Notation for the effective-UE product is inconsistent: κυ, κ v, κ̃υ and κ̃v appear interchangeably. A single symbol (e.g. κυ) should be fixed throughout the text, equations and figure captions.","section":"Throughout"},{"comment":"The abstract states that the analysis uses a “Gaussian mixture shadowing model”, yet the closed-form results rely on the mixture-exponential approximation. A short clarifying phrase would align the abstract with §II.B.","section":"Abstract"},{"comment":"Throughput values T3 ≈ 0.7 and T1 ≈ 0.9 bit/s/Hz are obtained by trapezoidal integration of the curves in Fig. 4. Reporting the numerical quadrature tolerance or supplying a short script would improve reproducibility.","section":"§IV"},{"comment":"Figure 1 caption and the surrounding text refer to a “projection into the line”; a brief sentence explaining that the figure is only schematic (the actual process lives on R^{2}) would help readers unfamiliar with the mapping.","section":"Fig. 1"}],"recommendation":"minor_revision","confidential_remarks":"The manuscript is a natural and technically clean extension of the authors’ earlier meta-distribution work [5]. The novelty relative to the terrestrial SIC literature and to the single LEO-SIC reference [2] is adequately disclosed. Scope fits a communications / stochastic-geometry journal; length is appropriate for a letter or short paper. No citation or ethical concerns."},"author_rebuttal":null,"desk_editor":{"model":"grok-4.5","letter":"This is the first explicit Poisson-Dirichlet (0,θ) treatment of ordered SIR under SIC for a narrow-beam LEO uplink. That is the real novelty: terrestrial SIC literature uses PD(α,0), and earlier LEO work did not give the factorial-moment densities or the SIC coverage integrals. The mapping to a gamma process (Laplace transform (12)), the densities in Prop. 2, the joint order-statistic PDF (18), and the SIC probability (25) are all rigorous once you accept the mixture-exponential shadowing. Spherical Monte-Carlo with the true 3GPP Gaussian mixture sits inside the analytic curves for 30°–90° elevation, so the final numbers are not free-floating.\n\nWhat they show is concrete: with κυ = n = 3 the third-UE coverage at –7 dB jumps roughly 900 % relative to the non-SIC k-probability, total throughput only falls from ~0.9 to ~0.7 bit/s/Hz, and the strongest-UE SIR variance stays finite. That is useful quantitative guidance for uplink multiplexing in commercial LEO systems.\n\nSoft spots are real but secondary. The mixture-exponential is fixed by matching the first two moments of the urban LoS/NLoS mixture; everything after that is exact only under that approximation. The paper never quantifies the higher-moment error, yet the MC envelopes already bound the practical discrepancy for the reported operating point. Perfect SIC and the planar narrow-beam idealization are modeling choices, not hidden contradictions. No code is released, but every parameter is tabulated, so re-implementation is straightforward. Self-citation to their earlier geometric constant κ is limited and re-derived.\n\nThis is for the stochastic-geometry satellite crowd and for system designers who need order-statistic numbers rather than another capacity bound. It is solid enough that a serious editor should send it to referees; the main revision request would be a short quantitative check of the fading approximation. I would cite the PD densities and the SIC integrals if I were working on LEO uplink multi-user analysis.","headline":"Clean first derivation of PD(0,θ) order statistics for narrow-beam LEO uplink SIC; the multi-UE fairness claim is backed by formulas and spherical MC, with the moment-matched fading approx as the main soft spot.","tokens_in":10497,"tokens_out":539,"would_cite":true,"duration_ms":5093,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"grok-4.5","headline":"SIC lets each LEO satellite beam serve three users at once with only a modest throughput trade-off and finite SIR variance.","keywords":["low Earth orbit","stochastic geometry","successive interference cancellation","Poisson–Dirichlet","SIR order statistics","narrow-beam uplink","mixture shadowing"],"falsifier":"Re-run the Monte-Carlo campaign of Section IV with the original two-tier log-normal mixture (instead of the moment-matched exponential) and check whether the third-user SIC coverage at −7 dB still improves by roughly an order of magnitude and whether total throughput remains near 0.7 bit/s/Hz.","tokens_in":10456,"feed_emoji":"📡","tokens_out":676,"duration_ms":5730,"temperature":0.7,"pith_summary":"The paper asks whether successive interference cancellation can turn the natural ordering of user powers inside a narrow LEO beam into a practical multi-user uplink. Because the received-power process under mixture-exponential shadowing is a gamma process, the ordered signal-to-total-interference ratios follow a Poisson–Dirichlet law whose density is known in closed form. From that density the authors obtain exact coverage probabilities for the first, second and third strongest users both with and without SIC. Numerical evaluation for three effective users per beam shows that SIC raises the third user’s success probability at −7 dB by roughly nine hundred percent, keeps the total spectral efficiency near 0.7 bit/s/Hz, and places the system in the finite-variance regime of the SIR. The result matters because it supplies a concrete operating point at which a LEO beam can serve several users fairly without collapsing average throughput.","feed_headline":"SIC lets one LEO beam serve three users with only 0.2 bit trade-off","feed_subtitle":"Coverage of the weakest user jumps ~900 % while total throughput stays near 0.7 bit/s/Hz and SIR variance stays finite.","key_machinery":"The Poisson–Dirichlet density of the ordered STIR process (Proposition 2), obtained because the Laplace transform of total received power is that of a gamma random variable with shape κ̃υ; all subsequent SIC coverage formulas are obtained by integrating this density over the successive-decoding region.","core_discovery":"Under the mixture-exponential shadowing model the ordered STIR process at a typical LEO beam is Poisson–Dirichlet PD(0, κ̃υ). Consequently the SIC coverage probability of the k-th strongest user admits an explicit multi-dimensional integral over the joint density of the first k order statistics; evaluating that integral for κ̃υ log 2 = 3 shows that three users can be served simultaneously with total throughput only 0.2 bit/s/Hz below the single-user baseline while the strongest-user SIR variance remains finite.","pith_inferences":[],"forward_implications":[],"fun_headline_variants":["SIC serves three UEs per LEO beam at 0.2 bit/s/Hz total cost","One LEO beam covers three ordered users via SIC near baseline rate","SIC enables multi-UE LEO service with finite SIR variance and fairness","Three concurrent users per narrow LEO beam under SIC order stats","LEO SIC covers weakest UE while total throughput stays near 0.7 bit"],"cache_read_input_tokens":128,"weakest_assumption_plain":"The two-tier log-normal shadowing of 3GPP is replaced by a one-parameter mixture-exponential distribution whose parameter is fixed solely by matching the first two moments of received power.","fun_headline_variants_meta":{"raw":{"variants":["SIC serves three UEs per LEO beam at 0.2 bit/s/Hz total cost","One LEO beam covers three ordered users via SIC near baseline rate","SIC enables multi-UE LEO service with finite SIR variance and fairness","Three concurrent users per narrow LEO beam under SIC order stats","LEO SIC covers weakest UE while total throughput stays near 0.7 bit"]},"model":"grok-4.5","effort":"low","cost_usd":0.004068,"raw_usage":{"total_tokens":1168,"prompt_tokens":687,"num_sources_used":0,"completion_tokens":104,"cost_in_usd_ticks":40680000,"prompt_tokens_details":{"text_tokens":687,"audio_tokens":0,"image_tokens":0,"cached_tokens":128},"completion_tokens_details":{"audio_tokens":0,"reasoning_tokens":377,"accepted_prediction_tokens":0,"rejected_prediction_tokens":0}},"tokens_in":687,"tokens_out":104,"duration_ms":4363,"temperature":1.0,"reasoning_tokens":377,"cache_read_input_tokens":128,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-07-14T04:19:46.581030+00:00","model_set":{"reader":"grok-4.5"},"falsifier":"Re-run the Monte-Carlo campaign of Section IV with the original two-tier log-normal mixture (instead of the moment-matched exponential) and check whether the third-user SIC coverage at −7 dB still improves by roughly an order of magnitude and whether total throughput remains near 0.7 bit/s/Hz.","supporting_citations":[],"review_version":1}