{"id":"45adfe15-00d5-477a-bebd-f598915b9764","arxiv_id":"2608.08501","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"A survey that organizes hybrid beamforming research for LEO satellites and UAVs into a common five-category taxonomy and a platform-aware framework.","lead":"This paper surveys how hybrid analog-digital beamforming is being adapted for satellites and drones, organizing the existing research into five shared categories. It is a reference map for engineers and researchers working on non-terrestrial networks, not a new technical result.","discovery_kind":"review","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Reference-selection bias undermines the synthesis: the survey's central contrasts rest on an undocumented literature search.","rationale":"The reader's CONDITIONAL verdict identifies the absence of a literature-selection methodology as the weakest assumption. I agree. The survey's central novelty claim, 'no existing survey jointly reviews HBF for LEO satellite and UAV communication systems' (Section I-A), is a claim about the survey literature and could be checked by a targeted search, but the more consequential issue is the body of the survey: the five-category taxonomy and the recurring platform contrasts are only as good as the reference base. The paper does not document how the 100+ cited technical papers were found, screened, or chosen, so the synthesis may reflect the authors' reading rather than the field's actual structure. This is not an accusation of bias; it is a structural weakness: any survey's conclusions about relative emphasis, open problems, and cross-platform differences depend on the representativeness of the sample. The reader's recommended remedy, documenting the search protocol or softening the 'systematic review' claim, is exactly right. The paper otherwise appears well organized, with a coherent taxonomy, a useful comparison of architectures, and honest acknowledgement that most results are simulation-based with no common benchmarks. Therefore, I do not see a reason to move beyond the reader's conditional acceptance; the concern is real but addressable, so the verdict should remain CONDITIONAL as issued, meaning no change from the reader's assessment.","tokens_in":70,"tokens_out":3202,"duration_ms":50062,"concrete_test":"Reproduce the literature search systematically: query IEEE Xplore, Scopus, and Google Scholar for 2015–2026 with the Boolean expression (\"hybrid beamforming\" OR \"hybrid precoding\") AND (\"LEO\" OR \"low Earth orbit\" OR \"satellite\") AND (\"UAV\" OR \"unmanned aerial vehicle\" OR \"aerial\"), apply explicit inclusion criteria (e.g., peer-reviewed, proposes or analyzes an HBF architecture for LEO or UAV), and compare the resulting corpus with the paper's Tables II–IV. If the search surfaces more than 20% additional relevant papers that do not fit any of the five categories, or if beam-hopping appears in a substantial subset of UAV papers or beam-tracking in LEO papers, the synthesis is biased and the central contrast fails.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The paper's contribution is the claim to be the first survey that jointly reviews HBF for LEO and UAV under a common five-category framework (Section I-A). The value of that synthesis depends on the surveyed literature being representative of the two fields. Yet the paper provides no search strategy, database list, date range, or inclusion/exclusion criteria for the works in Tables II–IV. Several entries are the authors' own papers (e.g., [15], [16], [85], [132]) or closely related works, which may over-weight certain subtopics. If the reference base is not comprehensive, the sharpest contrast in the survey, traffic-driven beam hopping for LEO (Section III-B) versus mobility-aware beam tracking for UAV (Section IV-B), could be an artifact of selective reading rather than a true property of the literature. This is load-bearing because the paper's stated gap-filling contribution (Section I-A) and its cross-platform lessons (Sections II-F and VI) are general claims about the field, not just about the cited papers. Without a documented selection procedure, the reader cannot tell whether the taxonomy emerged from the literature or was imposed on it.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"This survey reviews hybrid analog-digital beamforming (HBF) for low-Earth-orbit satellite and UAV communication systems, which it treats as the two primary non-terrestrial platform axes. It first presents common HBF foundations (signal model, hardware architectures, channel models, analog and digital precoding, and learning-aided methods), then organizes the LEO and UAV literatures under a single five-category taxonomy: system architecture and precoding design, time-varying beam management, network-level design, sensing and reconfigurable surfaces, and security and multiple access. The central claim is that no existing survey jointly covers both platforms under such a common framework, and that the sharpest platform contrast lies in beam hopping for LEO versus mobility-aware beam tracking for UAVs. The paper concludes with cross-platform lessons and a set of open challenges in hardware-aware, wideband, ISAC, 3D multi-layer, and THz HBF design.","tokens_in":48308,"tokens_out":5629,"duration_ms":66545,"significance":"If the surveyed literature is representative, the paper provides a genuinely useful unified reference: the five-category taxonomy enables category-by-category comparison, and the beam-hopping-versus-beam-tracking contrast is an insightful organizing observation. The paper also makes good use of summary tables and architectural figures to map a large number of references to design objectives and platform couplings, and it explicitly connects simulation-level results to hardware-implementation concerns such as predistortion, calibration, and true-time-delay architectures. However, the value of the synthesis is contingent on the representativeness of the selected references in Tables II-IV, and the manuscript currently provides no methodology for how those references were found, screened, or coded. This is the main correctness risk for a survey whose contribution is the synthesis itself.","major_comments":[{"comment":"The survey's central claim and its main cross-platform conclusions depend on the representativeness of the references in Tables III and IV, but the paper provides no methodology for selecting or screening the literature. The abstract promises a 'systematic review,' yet there is no search strategy, no list of databases, no date range, and no inclusion or exclusion criteria; Section I-A asserts the research gap ('no existing survey jointly reviews...') and Section II-F draws general cross-platform lessons without explaining how the cited set was assembled. Because the sharpest contrast, traffic-driven beam hopping for LEO (Section III-B) versus mobility-aware beam tracking for UAVs (Section IV-B), is derived from the selected works, an undocumented selection process makes it impossible to tell whether this contrast is a property of the field or an artifact of the sample. This is load-bearing for a survey whose contribution is the synthesis. Please add a reproducible methodology subsection describing the search, screening, and coding procedure, and use it to audit Tables III and IV for coverage and balance; also report the number of candidate papers screened and the number and reasons for exclusions.","section":"I-A, I-B, Tables III-IV"},{"comment":"The five-category taxonomy is presented as the organizing principle of the survey, but no operational criteria are given for assigning a paper to a category, and no evidence is provided that the five categories are exhaustive for the LEO and UAV HBF literature. For example, 'sensing capability and reconfigurable surfaces' combines ISAC, RIS/RHS, lens, and holographic architectures, while 'security and multiple access' groups at least three distinct problem families; whether these groupings emerged from the literature or were imposed a priori is not justified. This matters because the paper's contribution is the claim of a common framework that permits category-by-category comparison. Please define each category operationally (for instance, in terms of the primary objective or hardware constraint used for coding) and state how borderline or multi-topic papers were assigned.","section":"I-B and Tables I-IV"}],"minor_comments":[{"comment":"The rate expression in Eq. (3) does not explicitly state that b_k denotes the k-th column of B and that the digital precoder dimension requires N_RF >= K; please add this clarification to avoid ambiguity.","section":"II-A1, Eq. (3)"},{"comment":"The statement that 'beam-coherence time grows with beamwidth and range but shrinks with transverse speed' is supported by a vehicular-channel reference [92]; since the paper is about UAV links, please either add a UAV-specific reference or an explicit argument for why the vehicular result carries over.","section":"IV-B"},{"comment":"The claim of 'approximately 11-ms latency in [127]' should be accompanied by the specific measurement context reported in that paper, including the hardware platform, so that readers can judge whether the number is representative of onboard NTN processors.","section":"V-A3"},{"comment":"The text contains inconsistent spacing artifacts in 'UA V' and 'UA Vs' (for instance, in the abstract and Section IV headers); please harmonize these to 'UAV' in the final version.","section":"Throughout"}],"recommendation":"major_revision","confidential_remarks":"The manuscript fits the scope of IEEE Communications Surveys and Tutorials, and the proposed unified taxonomy is a reasonable basis for a survey. My main concern is that the reference selection is not documented, which weakens the completeness and representativeness claims that are central to the paper's novelty. The reference list also contains a noticeable number of 2025-2026 'early access' and arXiv entries, as well as several of the authors' own works; a documented search protocol and a re-audit of Tables III-IV would address this. I do not see a need to reject, but the revision should treat the methodology as a substantive requirement rather than a cosmetic addition."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Short version: this survey earns its keep, and the thing that makes it useful is the thing the authors claim — a single five-category taxonomy applied to both LEO satellite and UAV hybrid beamforming. I have not seen those two literatures aligned category-by-category before, and the organizing insight (LEO couples to quantities the payload can predict but not change; UAV couples to quantities the platform controls) is genuinely clarifying rather than decorative. Tables III and IV make the comparison usable, and the Section II primer gives a newcomer the shared vocabulary. The survey also has the decency to say plainly that the field is simulation-heavy, single-platform, and short on common benchmarks — that honesty is earned, and it is rare in surveys.\n\nThe main weakness is exactly what the reader flagged: no documented search protocol. No databases, no date range, no inclusion or exclusion criteria, so \"systematic review\" in the abstract overpromises relative to the method. That is a real gap but it is fixable — a methods paragraph and a softened claim would do most of the work. The stress-test worry that the headline contrast (beam hopping on LEO vs beam tracking on UAV) is an artifact of selection does not land as hard, though. The contrast is well motivated by platform physics — deterministic orbital motion vs jitter and controllable trajectory — and the cited papers genuinely do cluster that way. Selection bias is a threat to the comprehensiveness claims; I am less convinced it undermines the central taxonomy.\n\nThe self-citations ([15], [16], [85], [132]) are visible but diluted over roughly 165 references, and the most prominent ones are honestly on-topic. Minor irritants: a ResearchSquare preprint sits in the reference list, a few venues look obscure, and the \"first joint survey\" claim is plausible but not verifiable from Table I alone.\n\nWho is this for: graduate students and engineers entering NTN beamforming, who need a map of both literatures and a sense of where the real hardware constraints bite. It deserves a serious referee — conditional acceptance asking for the search methodology would be a fair outcome, and with that addition this would be a genuinely useful COMST contribution.","headline":"A useful first joint LEO-UAV hybrid beamforming survey whose five-category taxonomy holds up; the undocumented literature search is the one real, fixable flaw.","tokens_in":48870,"tokens_out":3442,"would_cite":true,"duration_ms":38247,"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":"This survey claims to be the first to review hybrid beamforming for LEO satellites and drones under one five-category framework that makes the two platforms directly comparable.","keywords":["hybrid beamforming","non-terrestrial networks","LEO satellite","UAV","massive MIMO","mmWave","ISAC","beam hopping"],"falsifier":"Run a documented systematic search, with named databases, explicit inclusion and exclusion criteria, and a defined date range up to the paper's 2026 submission, for any survey that jointly reviews hybrid beamforming for both LEO satellite and UAV systems; finding one published before this paper refutes the stated gap. Independently, re-derive the paper's central contrast, beam hopping for LEO versus beam tracking for UAV, from the complete retrieved corpus rather than from Tables II–IV; if the literature outside the selection organizes differently, the taxonomy's generality weakens.","tokens_in":47904,"feed_emoji":"🛰️","tokens_out":12210,"duration_ms":99915,"temperature":0.7,"pith_summary":"Hybrid analog–digital beamforming (HBF) lets a large antenna array be driven by only a few radio-frequency chains, which makes it a key physical-layer technology for low-Earth-orbit satellites and UAVs, whose payloads and batteries cannot support fully digital arrays. This survey claims to be the first to review HBF for both platform families together, under a single five-category taxonomy that lets LEO satellite and UAV designs be compared category by category. Its central lesson is that the beamforming architecture must be co-designed with the platform: LEO payloads must cope with orbital motion, Doppler, and channel aging they can predict but not change, while UAVs must co-design the beam with altitude, position, and trajectory they control. The payoff for a reader is a shared map of which HBF problems are common to both platforms and which differ, together with an account of what still blocks real deployment.","feed_headline":"One taxonomy now unifies LEO and drone hybrid beamforming","feed_subtitle":"A five-category map of LEO and drone beamforming shows what each platform must co-design.","key_machinery":"The load-bearing structure is the five-category taxonomy applied identically to LEO and UAV systems, built on the shared HBF signal model $x = FBs$, where the constant-modulus analog precoder $F$ maps $N_{\\mathrm{RF}}$ radio-frequency chains to $N_t$ antennas and the low-dimensional digital precoder $B$ operates on the reduced effective channel $\\tilde{H} = HF$. Two further mechanisms carry the argument. The first is multi-scale temporal decoupling: on an LEO payload, analog beam-steering coefficients need updates only on a seconds scale while digital precoding must be refreshed at millisecond intervals, a split that makes hybrid architectures viable under tight power budgets. The second is the platform-coupling distinction: LEO robustness is against CSI aging, Doppler, and beam squint, whereas UAV robustness is against jitter, trajectory variation, and blockage, which is why the same five categories fill in so differently for the two platforms.","core_discovery":"The paper asserts that no existing survey jointly reviews hybrid beamforming for LEO satellite and UAV communication systems under a common platform-aware, architecture-aware, and methodology-aware framework, and that this survey fills that gap. It organizes both literatures under one set of five categories, namely system architecture and precoding design, time-varying beam management, network-level design, sensing and reconfigurable surfaces, and security and multiple access, and shows that the categories fill in differently in a revealing way: the dominant time-varying mechanism is traffic-driven beam hopping on an LEO payload but mobility-aware beam tracking on a UAV. The survey's own lesson is that architecture and algorithm must be co-designed with the platform, not selected after it: LEO couples to quantities the payload can predict but cannot alter, such as orbital trajectory, Doppler, beam squint, and channel aging, which favors predicted or statistical CSI and a slow-analog/fast-digital update split, whereas UAV couples to quantities the platform controls, such as altitude, position, and trajectory, making placement a first-class beamforming variable.","pith_inferences":["The framework likely generalizes beyond the two named platforms: the survey itself notes that high-altitude platforms inherit the same mechanisms, and the five categories would map without much distortion onto any mobile large-array link, so the taxonomy's usefulness may outlive its two home platforms.","A testable prediction the paper does not make itself: if the taxonomy is as natural as claimed, new NTN beamforming papers outside its selection should slot into the five categories with little forcing.","The reported update-rate split, seconds-scale analog versus millisecond-scale digital for LEO, implies a design rule the paper leaves implicit: implementations should budget analog and digital refresh separately rather than merge them into one beam-tracking loop.","The survey concedes that most reviewed designs are simulation-only; a natural follow-up it does not propose is hardware-in-the-loop replication of its headline contrast, beam hopping on a real payload front end versus jitter-robust tracking on a real drone array."],"forward_implications":["LEO and UAV HBF work can now be compared directly: the same five categories organize both literatures, so a choice such as fully-connected versus partially-connected arrays is judged by the same yardstick on a satellite and on a drone.","Exploiting the analog/digital update-rate split, with analog steering refreshed on a seconds scale and digital precoding at millisecond intervals, can sharply cut onboard computation on LEO payloads without sacrificing sum rate.","CSI quality outweighs CSI quantity: for both platforms, location, ephemeris, sensing, and statistical CSI can outperform delayed instantaneous CSI, redirecting effort from feedback bandwidth toward prediction.","Wide-beam and rate-splitting schemes degrade gracefully under beam misalignment and imperfect CSI, making jitter-robust analog beams and RSMA natural robustness layers for NTN HBF.","Learning-aided HBF earns its place when anchored to physical structure, such as deterministic orbital motion, beamspace sparsity, or trajectory-driven channel evolution, rather than as a black-box replacement for the radio model."],"supporting_citations":[{"why":"Foundational HBF survey that supplies the architecture families, signal model, and notation the entire NTN review extends.","marker":"[3]"},{"why":"Satellite-only precoding survey; the paper defines its gap partly by what this digital-focused survey does not cover.","marker":"[28]"},{"why":"Broad NTN-evolution survey used to show that network-level surveys leave transceiver-level HBF out.","marker":"[1]"},{"why":"NTN modeling survey used to show that the analytical-coverage line also leaves HBF transceiver design out.","marker":"[2]"},{"why":"Anchors the LEO statistical-CSI precoding line; its fully and partially connected subarray analysis grounds the LEO architecture category.","marker":"[4]"},{"why":"Anchors the UAV massive-MIMO HBF line; its closed-form aerial base-station rate analysis grounds the UAV architecture category.","marker":"[8]"},{"why":"Quantifies seconds-scale analog versus millisecond-scale digital update rates on LEO, load-bearing for the multi-scale temporal decoupling lesson.","marker":"[36]"},{"why":"MAPPO-based joint beamforming and beam hopping, load-bearing for the claim that LEO's time-varying mechanism is traffic-driven beam hopping.","marker":"[14]"},{"why":"Jitter-robust wide-beam UAV design, load-bearing for the claim that the UAV time-varying mechanism is mobility- and jitter-driven beam tracking.","marker":"[11]"}],"fun_headline_variants":["Unified survey maps LEO and drone hybrid beamforming","One taxonomy ties LEO and drone beamforming design","LEO vs drones: Two beamforming worlds, one map","Survey unifies satellite and drone beamforming co-design","Beam hopping vs beam tracking: one survey covers both"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The whole synthesis stands on the selected papers in Tables II–IV being a fair sample of the field, yet the paper never says how it searched for, screened, or chose those papers, so a biased selection would make its claimed platform contrasts unreliable.","fun_headline_variants_meta":{"raw":{"variants":["Unified survey maps LEO and drone hybrid beamforming","One taxonomy ties LEO and drone beamforming design","LEO vs drones: Two beamforming worlds, one map","Survey unifies satellite and drone beamforming co-design","Beam hopping vs beam tracking: one survey covers both"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000411,"raw_usage":{"total_tokens":2177,"prompt_tokens":1042,"completion_tokens":1135,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":658,"completion_tokens_details":{"reasoning_tokens":1055}},"tokens_in":658,"tokens_out":1135,"duration_ms":8530,"temperature":1.0,"reasoning_tokens":1055,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T04:33:21.544753+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Run a documented systematic search, with named databases, explicit inclusion and exclusion criteria, and a defined date range up to the paper's 2026 submission, for any survey that jointly reviews hybrid beamforming for both LEO satellite and UAV systems; finding one published before this paper refutes the stated gap. Independently, re-derive the paper's central contrast, beam hopping for LEO versus beam tracking for UAV, from the complete retrieved corpus rather than from Tables II–IV; if the literature outside the selection organizes differently, the taxonomy's generality weakens.","supporting_citations":[{"cited_title":"Precoding for High Throughput Satellite Communication Systems: A Survey","cited_arxiv_id":"2208.08542","evidence_quote":"Satellite-only precoding survey; the paper defines its gap partly by what this digital-focused survey does not cover."}],"review_version":1}