{"id":"d08d0588-0a7a-4f2f-a4a5-77f0da634c14","arxiv_id":"2505.10306","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"RAA achieves a direction-independent angular resolution of arcsin(2/M) for a selected sULA, while conventional ULA resolution degrades away from boresight.","lead":"Ray antenna arrays spread many cheap antennas along angled lines, so a receiver can steer a beam by picking the right line instead of using expensive phase shifters. This paper shows the array keeps the same angular resolution in every direction, which standard linear arrays lose when targets are off to the side, and applies it to UAV-swarm sensing and communication.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Theorem 1 assumes a continuously rotatable sULA (η=θ′), while Eq. (18) fixes a discrete grid; off-grid targets see an asymmetric, partially attenuated beam, so the exact uniform-resolution claim is not established for the implemented RAA.","rationale":"The central claim is Theorem 1's exact uniform-resolution result, and the proof relies on η=θ′. The physical design in Eq. (18) only provides a discrete set of orientations, so the theorem as proved does not directly apply to arbitrary target directions. This is not cosmetic: an off-grid target is offset from the peak of every sULA by up to half the inter-beam spacing, causing up to about 3.9 dB single-beam gain loss and asymmetric null geometry. A narrow reading of the paper's Definition 1 keeps the null-to-null width constant for any θ′ inside a selected mainlobe, but that reading does not correspond to the usual two-target resolvability measure, for which the target position within the mainlobe matters. The proposed test—sweeping the target centroid across one inter-sULA interval in the actual MUSIC algorithm—will settle whether the practical resolution ripple is negligible or not. I do not see a reason to escalate beyond the reader's conditional verdict: the mathematical core is coherent under its idealization, the simulation study is plausible, and the RAA architecture still offers an advantage over ULA at large angles. However, the paper should either prove a discrete version of Theorem 1 with an explicit bound on the off-grid error or clearly state the uniform-resolution claim with the quantization effect made explicit.","tokens_in":18337,"tokens_out":18765,"duration_ms":201696,"concrete_test":"Analytically recompute the resolution for an off-grid direction: replace η=θ′ in Eq. (12) with the nearest discrete orientation η_n = argmin_{n}|θ′−η_n|, and compute the nulls bracketing θ′ and the array response at θ′ for θ′ = η_n + Δ/2. Then run Algorithm 1 for two equal-power targets whose centroid is swept continuously across one inter-sULA interval [η_n, η_{n+1}] (e.g., 0° to 0.895° for M=128) with a fixed angular separation near the nominal resolution, and record detection probability and RMSE versus the fractional offset. If detection probability or RMSE oscillates periodically and degrades near θ′ = η_n + Δ/2, the discrete RAA does not realize the uniform resolution claimed in Theorem 1.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The load-bearing step is the identification of the fixed-beam RAA with a continuously steerable array. Theorem 1 is proved by setting the sULA orientation η equal to the desired direction θ′ in Eqs. (8), (12), and (16), yielding first nulls at θ′±arcsin(2/M). The implemented RAA, however, has only the discrete orientations η_n = n·arcsin(2/M) fixed by Eq. (18). For a target at arbitrary θ′, the RSN can only select the nearest sULA, with offset δ = θ′−η_n satisfying 0<|δ|≤Δ/2, where Δ=arcsin(2/M). The actual beam pattern is r(θ,η_n)=M√G(θ−η_n)·H_M(sin(θ−η_n)), whose peak is at η_n, not at θ′. Consequently, the array response at θ′ is reduced by up to |H_M(sin(Δ/2))|≈2/π, i.e., about −3.9 dB, relative to the on-grid case, and the first nulls lie at η_n±Δ, which are asymmetric with respect to θ′: the distances to the two nulls are Δ−δ and Δ+δ. A pair of targets centered at θ′ is therefore not symmetrically placed inside the mainlobe; when δ approaches Δ/2, one member of the pair can fall at or beyond a null while the other is near the peak, so the minimum resolvable separation and the MUSIC detection probability depend on the fractional offset θ′/Δ. The paper never quantifies this off-grid behavior or proves that the uniform-resolution statement survives the discretization. Theorem 1 as stated is a statement about a hypothetical sULA that can be rotated to exactly θ′, not about the discrete RAA of Eq. (18) used in the simulations and algorithms.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper proposes the ray antenna array (RAA) as a cost-effective multi-antenna architecture for low-altitude UAV swarm ISAC. An RAA consists of a set of simple uniform linear arrays (sULAs) with distinct fixed orientations, each with its elements directly combined, plus a ray selection network (RSN) that connects a subset of sULAs to RF chains. The authors derive the array response model, analyze the beam pattern of a single sULA, and claim that RAA achieves direction-independent angular resolution γ_RAA(θ′) = arcsin(2/M) (Theorem 1). They compare this with the resolution of a conventional ULA (Theorems 2 and 3), concluding that RAA is never worse and is strictly better away from boresight. They then design an OFDM-ISAC sensing algorithm combining MUSIC-based AoA estimation, zero-forcing spatial filtering, and 2D Periodogram for delay/Doppler estimation, with simulations showing improved resolution and communication rate over ULA.","tokens_in":18778,"tokens_out":11038,"duration_ms":100794,"significance":"The uniform-angular-resolution property, if established for the actual discrete RAA, would be a genuinely useful contribution: it would provide a phase-shifter-free array that maintains good angular resolution for targets far from boresight, which is directly relevant to low-altitude UAV swarm sensing. The paper's analytical formulas for the sULA beam pattern and the ULA/RAA resolution comparison are clean and self-consistent under the stated idealizing assumptions. The proposed receive processing chain (MUSIC + ZF + 2D Periodogram) is complete and the simulation study is reasonably extensive. The strengths of the paper are its closed-form beam-pattern analysis for the idealized continuously rotatable sULA, the explicit comparison theorems for ULA, and the concrete algorithm and numerical demonstrations.","major_comments":[{"comment":"Theorem 1 is proved by setting the sULA orientation η exactly equal to the desired direction θ′ in Eqs. (8), (12), and (16). However, the designed RAA in Eq. (18) has only discrete orientations η_n = n arcsin(2/M). For a target at an arbitrary θ′, the RSN selects the nearest sULA with offset δ = θ′ − η_n, 0 < |δ| ≤ arcsin(2/M)/2. The response at θ′ is then reduced by up to |H_M(sin(δ_max))| ≈ 0.64 (about −3.9 dB for M = 128), and the first nulls are at θ′ − (γ+δ) and θ′ + (γ−δ), asymmetric about θ′. The paper does not quantify this off-grid behaviour or prove that the uniform-resolution claim holds for the discrete RAA used in the simulations and algorithms; this is a load-bearing gap because uniform resolution is the paper's headline contribution.","section":"§IV.A, Theorem 1 vs. Eq. (18)"},{"comment":"The assertion that RAA achieves higher beamforming gain than ULA by using directional antenna elements is stated to be 'rigorously proved in [38]', which is described as 'in preparation' and thus not available to the reader. Since higher beamforming gain is listed as one of the three main advantages of RAA and is used to explain the communication-rate results in Fig. 10, the proof should either be included in this paper or the claim should be explicitly presented as an assumption depending on the chosen element pattern models, rather than as an externally established fact.","section":"§IV.A, after Eq. (18)"},{"comment":"The derivation of d²γ_ULA/dx² is not correct as written. The paper writes dγ_ULA/dx = 4x/(MΔ) and then concludes d²γ_ULA/dx² = 4/(MΔ), but Δ is a function of x through u = x+2/M and v = x−2/M, so the second derivative must include a term involving dΔ/dx. The conclusion γ_ULA(θ′) ≥ γ_RAA may still be true, but the proof in the appendix is incomplete and should be replaced with a correct convexity argument (e.g., showing that the second derivative of arcsin is positive on the relevant interval).","section":"Appendix A, proof of Theorem 3"}],"minor_comments":[{"comment":"The MUSIC spectrum is written as P_MUSIC(θ) = 1/(h_s^H(θ) E_n E_n^H h_s(θ0)); the second factor h_s(θ0) appears to be a typo for h_s(θ), and the denominator should be the real quadratic form h_s^H(θ) E_n E_n^H h_s(θ).","section":"Eq. (38)"},{"comment":"The definition of angular resolution as half the null-to-null mainlobe width is well-defined for symmetric beams, but the off-grid sULA patterns discussed above are asymmetric about θ′. Please clarify whether the resolution metric is meant to describe the width of the selected sULA's mainlobe or the achievable resolution for targets centred at an off-grid direction, and discuss the impact of asymmetry.","section":"Definition 1 and §IV.A"},{"comment":"The definition of 'average missing shots' uses card(S_i^θ) − card(hat S_i^θ) but the condition card(S_i^θ) ≥ card(hat S_i^θ) is only stated to hold for small noise; please state explicitly what is counted when the estimated set is larger than the true set.","section":"Eq. (46)"},{"comment":"There are several typographical and formatting issues, including inconsistent spacing in 'UA V' and the use of 'specturm' in figure captions; a careful proofread is recommended.","section":"Throughout"},{"comment":"For the RAA 'directional' case the element gain G_0 is set to 5.13 dB while ULA uses 0 dB, with the same total power; this is consistent with the paper's model but should be clearly stated as the mechanism for the RAA gain advantage, since the figure alone does not separate the effect of array architecture from that of the chosen element patterns.","section":"§VI, Fig. 10"}],"recommendation":"major_revision","confidential_remarks":"The central theorem (Theorem 1) is proved for a hypothetical sULA that can be rotated to exactly the desired direction, but the implemented RAA uses a discrete grid of orientations; without an analysis of the off-grid case the main contribution is not fully established for the actual system. The use of the authors' own unpublished reference [38] for the beamforming-gain proof is a self-referential support issue that should be resolved in revision. The proof of Theorem 3 in Appendix A is also incorrect as written, though the statement may still be true. These issues are fixable within the scope of the manuscript, so I recommend major revision rather than rejection."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The thing to know: the paper's central claim is real but narrower than the abstract suggests. Theorem 1 proves uniform angular resolution for a single sULA whose orientation exactly equals the target direction. The implemented RAA (Eq. 18) uses discrete orientations spaced exactly by the null width, so an off-grid target sits off every sULA peak and suffers up to ~4 dB attenuation with asymmetric nulls. The paper never quantifies that degradation. That is the genuine soft spot.\n\nWhat is new and good: Theorem 1 is a clean, correct observation, and the comparison against ULA is fair. Theorem 2 is standard, and the convexity argument in Theorem 3 is sound. The RAA-specific MUSIC pipeline (ZF spatial filtering plus 2-D Periodogram for delay/Doppler) is a reasonable engineering adaptation, and the simulations match the qualitative predictions. The paper is careful to separate the resolution analysis from the beamforming-gain claim, though that claim relies on ref [38], which is unpublished. That is self-referential support, but it does not damage the main resolution theorem. No code, data, or error bars are included, so the curves are not independently checkable.\n\nThe math holds up under its assumptions. The weakness is that the headline claim is stated for an idealized continuous alignment, and the gap to the discrete implementation is left unquantified. That is a revision-level issue, not a reject-level one.\n\nI would bring it to a reading group only if the group cares about cost-efficient array architectures; otherwise it is a competent subfield paper. It deserves peer review, with a clear request to quantify the off-grid resolution loss and to either prove or properly support the beamforming-gain advantage.","headline":"A correct but idealized resolution theorem; the discrete-array gap needs quantifying before this is ready.","tokens_in":19237,"tokens_out":2307,"would_cite":false,"duration_ms":24801,"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":"Ray antenna arrays give uniform angular resolution in every direction, and the paper applies that property to low-altitude UAV-swarm ISAC.","keywords":["ray antenna array","uniform angular resolution","integrated sensing and communication","UAV swarm","OFDM ISAC","MUSIC angle estimation","beam pattern","hybrid beamforming"],"falsifier":"For $M=8$, take a target at $\\theta'=0.15$ rad between the two nearest ray orientations $\\eta_0=0$ and $\\eta_1=\\arcsin(2/8)\\approx 0.2527$ rad. The selected sULA's nulls lie at $\\theta=\\pm 0.2527$ rad, so the distances from the target to the left and right nulls are approximately 0.4027 rad and 0.1027 rad, not both equal to the claimed $\\gamma_{\\rm RAA}=0.2527$ rad; measuring the actual two-target resolution at such off-grid angles against Theorem 1 would settle whether the discrete RAA really has constant resolution.","tokens_in":18149,"feed_emoji":"📡","tokens_out":12587,"duration_ms":113218,"temperature":0.7,"pith_summary":"The paper aims to establish that a ray antenna array (RAA) — $N$ simple uniform linear arrays (sULAs) of $M$ directly connected elements arranged radially — resolves targets with the same angular accuracy $\\gamma_{\\rm RAA}=\\arcsin(2/M)$ no matter which direction they come from. If that holds, it removes the main blind spot of conventional uniform linear arrays, whose resolution degrades as a target moves toward endfire, and it does so without expensive phase shifters. The paper derives the beam pattern, designs the ray orientations and ray selection network, and gives a complete OFDM-ISAC sensing pipeline for low-altitude UAV swarms: MUSIC for angle-of-arrival, zero-forcing spatial filtering, and a 2-D Periodogram for delay and Doppler. Simulations against DFT-codebook hybrid beamforming show RAA keeping angle RMSE flat and communication rate higher while ULA fails as the swarm moves off-boresight.","feed_headline":"Uniform angular resolution in every direction, without phase shifters","feed_subtitle":"Radial sULA layout keeps tight beams for off-boresight UAV targets and cuts hardware cost.","key_machinery":"The load-bearing object is the Dirichlet kernel $H_M(x)=\\frac{1}{M}\\sum_{m=0}^{M-1}e^{j\\pi m x}=e^{j\\pi(M-1)x/2}\\frac{\\sin(\\pi Mx/2)}{M\\sin(\\pi x/2)}$, evaluated at $x=\\sin(\\theta-\\eta_n)$ for an sULA. Its zeros at $x=\\pm 2/M$ make the first nulls of an aligned sULA symmetric in angle, independent of $\\theta'$; in a ULA the same kernel appears at $\\sin\\theta-\\sin\\theta'$, so its zeros are not symmetric in angle. The ray selection network then chooses the strongest sULA outputs, so beamforming is replaced by spatial selection and no phase shifters are needed.","core_discovery":"With an sULA aligned exactly to the desired direction $\\theta'$, the paper proves (Theorem 1) that the RAA beam pattern is $M\\sqrt{G(\\theta-\\theta')}\\,|H_M(\\sin(\\theta-\\theta'))|$, whose first nulls are at $\\theta'\\pm\\arcsin(2/M)$; hence $\\gamma_{\\rm RAA}(\\theta')\\equiv\\arcsin(2/M)$. The paper contrasts this with a ULA, whose pattern has the Dirichlet kernel in $\\sin\\theta-\\sin\\theta'$, giving $\\gamma_{\\rm ULA}(\\theta')=\\frac{1}{2}[\\arcsin(\\sin\\theta'+2/M)-\\arcsin(\\sin\\theta'-2/M)]$, which grows with $|\\theta'|$. It then proves $\\gamma_{\\rm ULA}(\\theta')\\ge\\gamma_{\\rm RAA}$ with equality only at $\\theta'=0$, so RAA has no worse and generally better angular resolution. Because each sULA covers only a small angular sector, the array can also use more directional elements and obtain higher beamforming gain, and the discrete orientation rule $\\eta_n=n\\arcsin(2/M)$ makes adjacent sULAs orthogonal.","pith_inferences":["Editorial inference: with the discrete orientations actually implemented, the resolution should oscillate slightly as a target moves between two adjacent sULA orientations; computing that ripple and choosing whether extra rays are worth the added aperture is a natural follow-up.","Editorial inference: the same $\\sin(\\theta-\\theta')$ versus $\\sin\\theta-\\sin\\theta'$ distinction suggests that other geometries, such as arc or cylindrical arrays, may inherit a form of uniform resolution while trading aperture size.","Editorial inference: the energy-based ray selection assumes all sULA outputs can be swept to pick the strongest; a compressed or fast-switching selection could lower the sweep overhead without changing the resolution theorem.","Editorial inference: near-field UAV targets would break the plane-wave assumption behind the Dirichlet-kernel nulls, so testing whether uniform resolution survives in near-field ISAC is the next regime worth checking."],"forward_implications":["For a swarm near the array endfire, two targets separated by about $\\arcsin(2/M)$ (about $0.9^\\circ$ for $M=128$) remain resolvable with RAA, while a ULA of the same gain needs a larger separation and eventually cannot resolve them at all.","The communication rate no longer collapses for users far off-boresight, since the uniform beam pattern bounds inter-user interference in every direction rather than only near broadside.","The OFDM sensing pipeline (MUSIC angle estimation, zero-forcing spatial filtering, and 2-D Periodogram) gives separated delay-Doppler maps for each UAV with only $N_{\\rm RF}\\ll N$ radio-frequency chains.","Hardware cost drops because each sULA's elements are summed directly and the RSN uses switches instead of phase shifters; the paper quotes a commercial phase-shifter-switch example where RAA costs less than one percent of hybrid beamforming.","The price is physical size: RAA needs more antenna elements and a larger aperture than ULA, but the extra size is less problematic at millimeter-wave and terahertz wavelengths."],"supporting_citations":[{"why":"Introduces the ray antenna array architecture, the directly summed sULAs, and the orientation rule $\\eta_n=n\\arcsin(2/M)$ that Theorem 1 and the RSN design build on.","marker":"[37]"},{"why":"Supplies the proof that sULAs covering narrow angular sectors can use higher-directivity elements, the basis for the beamforming-gain advantage in Section IV.","marker":"[38]"},{"why":"Provides the MIMO-OFDM ISAC waveform and signal model that the delay-Doppler estimation pipeline is based on.","marker":"[10]"},{"why":"Gives the 3GPP antenna element radiation pattern used for the numerical beam patterns and rate/resolution simulations.","marker":"[39]"}],"fun_headline_variants":["Ray antenna array: equal beam sharpness everywhere, no phase shifters needed","Cheaper antenna array keeps beams sharp for off-boresight UAV targets","Ray array gives uniform angular resolution for UAV swarm sensing and comms","Uniform resolution in all directions with cost-effective ray antenna array","Ray antenna design delivers constant beam resolution for drone swarms"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The constant-resolution formula assumes a target direction can always be matched exactly by the orientation of some sULA; the implemented RAA has only finitely many discrete orientations, so off-grid targets see an asymmetric main lobe whose resolution error the paper does not quantify.","fun_headline_variants_meta":{"raw":{"variants":["Ray antenna array: equal beam sharpness everywhere, no phase shifters needed","Cheaper antenna array keeps beams sharp for off-boresight UAV targets","Ray array gives uniform angular resolution for UAV swarm sensing and comms","Uniform resolution in all directions with cost-effective ray antenna array","Ray antenna design delivers constant beam resolution for drone swarms"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000779,"raw_usage":{"total_tokens":3532,"prompt_tokens":1123,"completion_tokens":2409,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":739,"completion_tokens_details":{"reasoning_tokens":2333}},"tokens_in":739,"tokens_out":2409,"duration_ms":17641,"temperature":1.0,"reasoning_tokens":2333,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-15T21:12:09.164796+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"For $M=8$, take a target at $\\theta'=0.15$ rad between the two nearest ray orientations $\\eta_0=0$ and $\\eta_1=\\arcsin(2/8)\\approx 0.2527$ rad. The selected sULA's nulls lie at $\\theta=\\pm 0.2527$ rad, so the distances from the target to the left and right nulls are approximately 0.4027 rad and 0.1027 rad, not both equal to the claimed $\\gamma_{\\rm RAA}=0.2527$ rad; measuring the actual two-target resolution at such off-grid angles against Theorem 1 would settle whether the discrete RAA really has constant resolution.","supporting_citations":[{"cited_title":"Ray antenna array: A novel cost- effective multi-antenna architecture for enhanced wireless communica- tion,","cited_arxiv_id":null,"evidence_quote":"Introduces the ray antenna array architecture, the directly summed sULAs, and the orientation rule $\\eta_n=n\\arcsin(2/M)$ that Theorem 1 and the RSN design build on."},{"cited_title":"A novel cost-effective MIMO archi- tecture with ray antenna array for enhanced wireless communication performance,","cited_arxiv_id":null,"evidence_quote":"Supplies the proof that sULAs covering narrow angular sectors can use higher-directivity elements, the basis for the beamforming-gain advantage in Section IV."}],"review_version":1}