{"id":"86585725-5cbe-4b7e-859f-987ed8902c73","arxiv_id":"2607.09030","paper_version":1,"verdict":"CONDITIONAL","confidence":"HIGH","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":5,"one_line_summary":"A multi-satellite CAPA ICAN design reduces average navigation CRB under rate and power constraints by projecting continuous beamformers onto a joint channel subspace and solving an iterative SDP.","lead":"This paper designs continuous-aperture antennas on cooperating LEO satellites so they can send data and navigation signals on the same spectrum, then optimizes the continuous beam patterns to tighten positioning error while meeting rate targets. It matters because mega-constellations want dual-use payloads, and continuous apertures may beat discrete phased arrays on that trade-off.","discovery_kind":"new_application","skeptic_critique":{"model":"grok-4.5","headline":"The CRB objective used for optimization is not the true CRB once σ²_eff depends on the same beamformers that shape µl; the reported CAPA gains may partly be artifacts of that approximation.","rationale":"The reader correctly isolated the weakest modeling step: the weak-dependence treatment of σ²_eff. That step is not a minor numerical convenience; it is baked into both the performance metric that is optimized and the metric that is plotted. Theorem 1 and the subspace reduction remain valid, the SDP is correctly derived under the frozen-σ² model, and the CAPA-vs-DPA comparison is still informative under that model. Because the paper already flags the approximation and uses a damped update, the contribution stays accept-shaped for an ideal CAPA study, but the outperformance claim cannot be treated as operational until the full FIM (or empirical MSE) is checked. Hence the verdict remains CONDITIONAL; no stronger downgrade is warranted without the concrete test failing.","tokens_in":24042,"tokens_out":782,"duration_ms":8524,"concrete_test":"Re-evaluate the final beamformers of Algorithm 1 (and the DPA/Fourier baselines) with the full position-and-beamformer-dependent FIM that includes \nabla_q σ²_eff (or Monte-Carlo MSE of the DPE (28) with σ²_eff recomputed at every candidate q). If the CAPA-vs-DPA gap in average CRB/MSE shrinks by more than ~15–20 % relative to Figs. 4–6, or if the ranking reverses at low P_max or high R_min, the approximation is material to the claim.","verdict_should_be":"CONDITIONAL","load_bearing_attack":"The central claim is that the subspace-reduced CAPA design (Theorem 1 + Algorithm 1) yields substantially lower average navigation CRB than DPA/Fourier/ZF under identical power and rate constraints (Figs. 4–6). That claim rests on the FIM/CRB in (29)–(30) and the SDP objective (51)/(53). Both treat σ²_eff,l as independent of the optimization variables when forming the Jacobian and when inverting the FIM: after (24) the paper freezes σ²_eff inside each DPE/CRB evaluation, and (55) only damps it from the previous iterate. Yet (22)/(43)/(52) show σ²_eff,l = η\rho ∑_m |∑_k gN_k,l a_k,m|^{2} + σ^{2}_N,l, which is an explicit quadratic function of the same communication coefficients {A_m} that appear in the rate constraints and that couple into the navigation mean through residual leakage. When this dependence is non-negligible, the true FIM contains additional score terms from \nabla_q log σ²_eff and the optimized objective is no longer a valid upper bound on MSE; the ranking versus DPA and the absolute CRB numbers can therefore shift. The multi-satellite geometry and shared-spectrum ICAN setting make the residual term larger than in single-platform CAPA-ISAC, so the approximation is load-bearing for the headline outperformance result.","agreement_with_reader":"agree"},"referee_report":{"model":"grok-4.5","summary":"The paper develops a continuous-aperture array (CAPA) ICAN framework for multi-satellite LEO constellations. It models collaborative dual-function transmission via continuous surface currents and far-field dyadic Green’s functions, derives CUE rates and NUE CRBs under shared-spectrum interference, and formulates joint beamforming to minimize average CRB subject to per-satellite power and minimum-rate constraints. Theorem 1 shows that an optimal solution lies in a finite ICAN channel subspace spanned by the conjugate communication/navigation channel responses; the resulting finite-dimensional problem is solved by SDR, BCD, and SCA (Algorithm 1). Simulations on a Walker Delta constellation report lower average CRB than discrete phased-array, Fourier, ZF-oriented, and navigation-centric baselines under the same power and rate constraints.","tokens_in":24423,"tokens_out":717,"duration_ms":7544,"significance":"If the modeling and optimization claims hold, the work is a solid systems contribution at the intersection of electromagnetic information theory, LEO ICAN, and continuous-aperture beamforming. Strengths include an explicit multi-satellite CAPA model, a clean optimality-preserving subspace reduction (Theorem 1 / Appendix B), and a practical iterative SDP algorithm with convergence and complexity discussion. The numerical comparisons against DPA and other natural baselines under fixed Table I parameters make the performance claim falsifiable. The main novelty is the multi-satellite ICAN setting rather than a wholly new mathematical technique; the result is still of clear interest for 6G NTN dual-function design.","major_comments":[{"comment":"Section II-D after (24) and the FIM/CRB in (29)–(30) treat σ^{2}_eff,l as weakly dependent on ql and the beamformers, freezing or damped-updating it (III-C, (55)) while optimizing over {Am} and B. Yet (22)/(43)/(52) make σ^{2}_eff,l an explicit quadratic function of the same communication coefficients that enter the rate constraints and residual leakage. When residual communication interference is non-negligible (shared-spectrum multi-satellite ICAN), the true score includes ∇_q log σ^{2}_eff terms, so the optimized objective is not the exact CRB and absolute numbers / ranking versus DPA in Figs. 4–6 can shift. Please either (i) derive/optimize the full FIM including the variance dependence, or (ii) quantify the approximation error (e.g., relative contribution of residual terms and sensitivity of reported CRB gaps) under the operating points of Table I.","section":null},{"comment":"The headline claim that CAPA “significantly outperforms conventional discrete phased array architectures” (Abstract; §IV) rests on the DPA baseline in Figs. 4–6. The manuscript only briefly states that DPA uses the same aperture with half-wavelength spacing and optimized per-element weights [30]. Please specify the DPA element count, polarization model, and whether the same ICAN subspace / SDR-SCA solver (or an equivalent discrete formulation) is used, so that the gap is attributable to continuous aperture rather than unequal optimization effort or modeling assumptions.","section":null}],"minor_comments":[],"recommendation":"major_revision","confidential_remarks":null},"author_rebuttal":null,"desk_editor":{"model":"grok-4.5","letter":"The useful delta here is the multi-satellite LEO ICAN setting: cooperative CAPAs radiating shared-spectrum data streams plus per-satellite navigation PRNs, with rates and CRBs derived from the same continuous currents, plus a joint CUE–NUE channel subspace that turns the infinite-dimensional design into a finite SDP without losing optimality (Theorem 1 / Appendix B). That is incremental relative to the single-platform CAPA multiuser and CAPA-ISAC line they cite, but it is a real systems step for NTN ICAN.\n\nWhat they do well is the electromagnetic scaffolding. Far-field dyadic Green’s functions, polarization, rain, the power integral, the SINR decomposition, and the stacked DPE observation model are standard and internally consistent. The orthogonal-complement argument is correct: anything outside the span of the conjugate channels does not affect rates or CRB and only burns power, so the reduction is not a redefinition of the objective. Algorithm 1 (BCD + SCA penalty + damped σ²_eff) is ordinary SDP engineering and converges in the reported plots. Simulations under a Walker-like constellation show CAPA beating DPA, Fourier, ZF-oriented, and sitting close to the navigation-centric lower bound across power, rate, and aperture sweeps. Self-citations are to prior CAPA work, not circular.\n\nSoft spots, in proportion. The weakest modeling choice is the one the stress-test flags: after (24) they treat σ²_eff as only weakly position/beamformer-dependent and freeze or damp-update it inside the FIM and the SDP objective, even though (22)/(43)/(52) make it an explicit quadratic in the same communication coefficients that appear in the rate constraints. In a multi-satellite shared-spectrum setup that residual can be non-negligible, so the optimized “CRB” is not always the true CRB and absolute numbers (and possibly ranking vs DPA) can shift. That is a common approximation in the literature, not a derivation error, but it is load-bearing for the outperformance claim and should be stress-tested or jointly differentiated. Everything else is the usual idealization stack: perfect continuous current control, far-field LoS, perfect CSI/ephemeris, Gauss-Legendre integrals, simulation-only. No hardware or non-ideal evidence.\n\nWho it is for: people already working on CAPA / holographic MIMO or LEO ICAN who want a multi-satellite dual-function formulation and a workable optimizer. Not a foundational result. I would send it to peer review; a serious referee can demand a sensitivity check on σ²_eff and clearer limits of the ideal CAPA model. Worth engaging if that is your subfield; not urgent otherwise.","headline":"Solid multi-satellite CAPA ICAN formulation with a clean optimality-preserving subspace reduction; the headline CRB gains are real under the paper’s model but rest on a standard (and here load-bearing) freeze of σ²_eff.","tokens_in":25087,"tokens_out":711,"would_cite":true,"duration_ms":8273,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"grok-4.5","headline":"Continuous aperture arrays on cooperating LEO satellites cut navigation error while meeting communication rates better than discrete phased arrays, via dual-function beamformers optimized in a finite channel subspace.","keywords":["continuous aperture array","integrated communication and navigation","LEO satellite constellation","electromagnetic information theory","Cramér-Rao bound","beamforming","6G","channel subspace"],"falsifier":"Re-optimize and re-evaluate the CRB (or run Monte-Carlo position MSE) when the effective noise variance is fully differentiated with respect to both position and the beamformer coefficients; check whether the CAPA advantage over discrete phased arrays under the same power and rate constraints shrinks or vanishes.","tokens_in":24885,"feed_emoji":"🛰️","tokens_out":1032,"duration_ms":20558,"temperature":0.7,"pith_summary":"This paper builds an integrated communication-and-navigation system for LEO constellations in which multiple satellites, each fitted with a continuous aperture array, radiate both user data and navigation reference signals on the same spectrum. It derives the users’ achievable rates and the Cramér–Rao bound on position error, making explicit how the continuous dual-function beamformers couple the two services. The design problem—minimize average navigation CRB subject to per-satellite power limits and minimum rate floors—is infinite-dimensional; the authors prove that optimal beamformers live in a finite “ICAN channel subspace” spanned by the conjugate channel responses, convert the problem into a tractable SDP, and solve it by an iterative convex algorithm. Simulations show the resulting CAPA design substantially outperforms same-size discrete phased arrays and several other baselines on positioning accuracy while still satisfying the rate constraints. A reader who expects mega-constellations to deliver both connectivity and high-precision PNT cares because the continuous-aperture approach claims more spatial degrees of freedom and better dual-function performance without enlarging the physical aperture or spectrum.","feed_headline":"CAPA satellites beat discrete arrays on joint nav and data","feed_subtitle":"A channel subspace turns infinite continuous beamformers into a finite design that outperforms phased arrays under rate floors.","key_machinery":"The ICAN channel subspace (Theorem 1): the finite-dimensional span, on each satellite aperture, of the conjugates of the continuous communication and navigation channel responses. Restricting the continuous beamformers to this subspace leaves rates, navigation means, CRBs and power unchanged or improved, converting the original infinite-dimensional functional design into a finite-dimensional SDP solved by block-coordinate descent and successive convex approximation.","core_discovery":"Equipping a cooperative group of LEO satellites with continuous aperture arrays and jointly designing their dual-function continuous beamformers—after an optimality-preserving projection onto the finite-dimensional ICAN channel subspace—yields a lower average navigation CRB than conventional discrete phased-array architectures under identical per-satellite power budgets and minimum communication-rate constraints.","pith_inferences":["The same channel-subspace reduction is likely reusable for other multi-platform continuous-aperture dual-function problems (e.g., multi-satellite ISAC) that share the linear integral structure of the observation model.","If the weak-dependence approximation for effective noise variance fails under dense multi-user interference, jointly differentiating the CRB with respect to both mean and variance could alter the optimized beamformers and the reported gains.","Any practical CAPA realization (continuous current control or dense metasurface) that cannot match the idealized continuous current distribution will erode the simulated advantage over discrete arrays.","Perfect real-time CSI and ephemeris exchange over inter-satellite links is assumed; latency or estimation error would couple into both rate and CRB and remains an open implementation gap."],"forward_implications":["CAPA-based multi-satellite ICAN can achieve lower average navigation CRB than same-aperture discrete phased arrays while still meeting CUE rate floors.","Adding more cooperating satellites in the service group further reduces average CRB through spatial diversity and extra design degrees of freedom.","Larger CAPA area and denser or lower-altitude constellations improve positioning accuracy, with diminishing returns that should guide payload size and constellation density.","The explicit rate–CRB trade-off is tunable: raising the minimum rate increases CRB, yet the CAPA design degrades more slowly than zero-forcing or discrete-array baselines."],"fun_headline_variants":["CAPA LEO sats cut average nav CRB vs discrete arrays under rate floors","Continuous aperture beamformers beat phased arrays on LEO ICAN CRB","Subspace CAPA design lowers navigation CRB for joint LEO satellite ICAN","CAPA-assisted dual beamformers outperform discrete arrays in LEO ICAN","Joint CAPA beamforming yields tighter CRB than phased arrays under power and rate limits"],"cache_read_input_tokens":16512,"weakest_assumption_plain":"The navigation error bound treats the effective interference-plus-noise variance as only weakly dependent on user position and beamformers, so that variance is held fixed or lightly damped-updated rather than differentiated jointly with the signal mean.","fun_headline_variants_meta":{"raw":{"variants":["CAPA LEO sats cut average nav CRB vs discrete arrays under rate floors","Continuous aperture beamformers beat phased arrays on LEO ICAN CRB","Subspace CAPA design lowers navigation CRB for joint LEO satellite ICAN","CAPA-assisted dual beamformers outperform discrete arrays in LEO ICAN","Joint CAPA beamforming yields tighter CRB than phased arrays under power and rate limits"]},"model":"grok-4.5","effort":"low","cost_usd":0.004162,"raw_usage":{"total_tokens":1261,"prompt_tokens":755,"num_sources_used":0,"completion_tokens":108,"cost_in_usd_ticks":41620000,"prompt_tokens_details":{"text_tokens":755,"audio_tokens":0,"image_tokens":0,"cached_tokens":256},"completion_tokens_details":{"audio_tokens":0,"reasoning_tokens":398,"accepted_prediction_tokens":0,"rejected_prediction_tokens":0}},"tokens_in":755,"tokens_out":108,"duration_ms":5442,"temperature":1.0,"reasoning_tokens":398,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-07-13T00:52:13.260911+00:00","model_set":{"reader":"grok-4.5"},"falsifier":"Re-optimize and re-evaluate the CRB (or run Monte-Carlo position MSE) when the effective noise variance is fully differentiated with respect to both position and the beamformer coefficients; check whether the CAPA advantage over discrete phased arrays under the same power and rate constraints shrinks or vanishes.","supporting_citations":[],"review_version":1}