{"id":"d8874e17-ce76-4588-906b-3f5db6530945","arxiv_id":"2505.15200","paper_version":1,"verdict":"REJECT","confidence":"HIGH","novelty_score":4.0,"correctness_risk":"high","formal_verification":"none","parameter_count":1,"one_line_summary":"The paper gives outage and rate formulas for a single fluid antenna with N correlated ports under Rician fading, claims diversity order N, and says planar port layouts beat linear ones.","lead":"This paper analyzes a 'fluid antenna' receiver that switches among N positions on a small surface under Rician fading, deriving formulas for connection outage and data rate. The generalist takeaway: the authors claim more switchable positions improve reliability, but the key mathematical step may not be exact.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The 'exact' joint PDF in Lemma 1 (Eq. 12) is not exact for Rician fading: conditioning on |h1|=m1 does not remove the phase dependence of h1 in the conditional law of |hn|, so Eq. (15) inherits an unverified approximation.","rationale":"The reader's weakest_assumption is the same concern I find load-bearing: Appendix A's Eq. (A.2) treats the conditional law of |h_n| given |h_1| as a single Rician law with a phase-independent noncentrality. Direct calculation from Eq. (10) shows the conditional mean of h_n given h_1 is rho_n h_1 + (1-rho_n)A, whose magnitude depends on arg(h_1) through 2 rho_n (1-rho_n) m_1 A cos theta. Averaging over the conditional phase of h_1, which is not uniform when A != 0, does not yield the Rician form in Eq. (A.2). This is an internal inconsistency between Eq. (10) and Eq. (12), not a disagreement with consensus. For kappa=0, A=0, the phase-dependent term vanishes and Lemma 1 is exact, explaining why the formulas match the Rayleigh special case; for kappa>0 the advertised exactness fails. I also note a separate issue in Proposition 1: its proof treats |rho| = max_n |rho_n| as fixed while N tends to infinity, but Eq. (9) drives rho_2 to 1, so the simplified bound in Eq. (25) tends to 1 and the proof does not establish the claimed infinite slope. This second concern is independent and secondary. The paper contains no machine-checked proof or reproducible code, and the simulations reuse the same formulas, so they do not provide independent support for Eq. (15). Credit is due for the correct Rayleigh special case and for a reasonable simulation methodology, but these do not rescue the central claim of exact Rician outage analysis; rejection is appropriate, with the possibility that a corrected phase-aware derivation could turn the result into a valid approximation.","tokens_in":20869,"tokens_out":10217,"duration_ms":86398,"concrete_test":"Run a Monte Carlo simulation of the exact channel model in Eq. (10) with kappa=2, W=1, N=4, and gamma_th=1 (0 dB): draw 10^7 realizations of x0,y0,xn,yn, form h_n, and estimate P(max_n |h_n|^2 <= gamma_th) together with its binomial standard error. Compare this empirical outage with the value predicted by Theorem 1, Eq. (15). At an outage near 0.1, the standard error is about 0.0095 percentage points, so a difference exceeding 3 sigma (about 0.03 percentage points) falsifies Eq. (12). As an analytic cross-check, numerically integrate the phase-averaged conditional Rician density with noncentrality |rho_n m_1 e^{j theta} + (1-rho_n)A| over the conditional phase density of h_1 and compare the resulting joint CDF with Eq. (13).","verdict_should_be":"REJECT","load_bearing_attack":"Under Eq. (10), h_n given h_1 follows CN(rho_n h_1 + (1-rho_n)A, sigma^2(1-rho_n^2)). Therefore, conditioned only on |h_1|=m_1, the distribution of |h_n| is a phase-average over theta=arg(h_1) of a Rician density whose noncentrality is |rho_n m_1 e^{j theta} + (1-rho_n)A| = sqrt(rho_n^2 m_1^2 + (1-rho_n)^2 A^2 + 2 rho_n (1-rho_n) m_1 A cos theta). The phase theta is not uniform conditional on m_1 when A>0; its density is proportional to exp(2 A m_1 cos theta / sigma^2). Eq. (A.2) omits the cos theta term and also uses (1-rho_n^2)A^2 instead of (1-rho_n)^2 A^2. Hence Eq. (12) is exact only for A=0 (kappa=0). For kappa>0 it is an approximation, and the 'exact' outage probability in Theorem 1 (Eq. 15), the bounds in Corollaries 1-4, the ergodic-rate expressions in Theorem 2 and Corollaries 5-6, and the UPA results all inherit that error. The Monte Carlo agreement in Figs. 3-5 does not validate exactness because the same formula is plotted as the analytical curve; a direct comparison against the true Eq. (10) is required.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"This manuscript analyzes a single-input single-output fluid antenna system (Rx-SISO-FAS) with N candidate ports under spatially correlated Rician fading. It proposes an exact joint PDF/CDF for the port amplitudes, an outage-probability expression, closed-form upper and lower bounds, ergodic-rate formulas, and diversity-order claims, for both ULA and UPA port layouts. The results are supported by Monte Carlo simulations and comparisons with fixed-antenna and MRC benchmarks. The paper's principal analytical object is Lemma 1, whose Eq. (12) is used to derive all subsequent outage and rate results.","tokens_in":21199,"tokens_out":12143,"duration_ms":102020,"significance":"If correct, the exact expressions would extend FAS performance analysis from Rayleigh and Nakagami-m models to Rician channels with arbitrary port correlation, and the UPA comparison would be a useful design insight. The manuscript is well organized, includes several useful reductions (the kappa=0 and N=1 cases limit to known results), and the ULA/UPA treatment is systematic. However, the central exactness claim is not supported: Lemma 1 is an approximation for kappa>0, and the phase dependence that is dropped propagates into Theorem 1 and the downstream results. The Monte Carlo agreement does not test the exact joint law, so the claimed exactness is not established.","major_comments":[{"comment":"The conditional density asserted in Eq. (A.2) is not a consequence of the channel model in Eq. (10). From Eq. (10), h_n given h_1 is CN(rho_n h_1 + (1-rho_n)A, sigma^2(1-rho_n^2)) for n>=2, so the Rician noncentrality parameter conditioned only on |h_1|=m_1 is sqrt(rho_n^2 m_1^2 + (1-rho_n)^2 A^2 + 2 rho_n (1-rho_n) m_1 A cos(theta)), where theta = arg(h_1), and theta is not uniform given m_1 when A>0: its conditional density is proportional to exp(2 A m_1 cos(theta)/sigma^2). Eq. (A.2) drops the cross term and uses (1-rho_n^2)A^2 in its place. Consequently Eq. (12) is exact only for A=0 (kappa=0); for kappa>0 it is an unstated approximation, not an identity.","section":"Appendix A / Lemma 1 (Eq. (12))"},{"comment":"Because Lemma 1 is the input to Lemma 2, the outage probability in Eq. (15), Corollaries 1-4, Theorem 2, Corollaries 5-6, and Theorem 3 for the UPA configuration all inherit the unverified approximation from Eq. (A.2). The curves labeled 'analytical' in Figs. 3-5 are computed from Eq. (15) or its corollaries, so agreement with Monte Carlo simulations at selected parameter points does not certify the exactness claim; a direct check would simulate the phase-preserving conditional law of Eq. (10) and compare the resulting empirical CDF with Eq. (15), or compute the phase-averaged integral of the true conditional Rician density numerically.","section":"Theorem 1 / Eq. (15) and downstream results"},{"comment":"The claimed lower bound on the ergodic rate is not proven. Eq. (E.1) is an upper bound on the CDF F(x), but the transition to Eq. (E.2) discards the factor (1-e^{-x}) and the terms involving kappa(1-rho_n^2) and 2 sqrt(kappa x(kappa+1)); dropping terms from an upper bound does not in general yield a lower bound, and no inequality is shown to justify the '>=' in Eq. (36). In addition, the constants a_i and b_i contain c, which is never assigned a value, so the expression is not fully specified as a closed-form result.","section":"Appendix E / Corollary 5 (Eqs. (E.1)-(E.2), (36))"}],"minor_comments":[{"comment":"The outage event is written as SNR_FAS/SNRn = |h_FAS|^2/(A^2+sigma^2) < gamma_th, but SNR_FAS/SNRn is |h_FAS|^2/|hn|^2 by the preceding definitions; the normalization used in the subsequent integrals should be defined consistently.","section":"Eq. (14)"},{"comment":"The series representation indexes n=1...N with rho_1=0, but the product over n=2...N in Eq. (16) is the starting point; please confirm that the n=1 term is the correct Rician CDF in this representation.","section":"Corollary 3 / Eq. (18)"},{"comment":"The ergodic-rate simulations use only 10^2 Monte Carlo samples, which is quite low for rate curves; reporting confidence intervals or increasing the number of samples would make the verification more convincing.","section":"Section V / Fig. 9"},{"comment":"The proof uses 'approximately equal' signs inside a limit calculation; a formal diversity-order statement should be phrased as a limit inferior/superior or with explicit asymptotic bounds.","section":"Proposition 2 / Eq. (28)"},{"comment":"There are minor typographical issues, including 'UP A' with a space in several places and 'the OP' at the start of Theorem 3; these should be corrected in a revision.","section":"Throughout"}],"recommendation":"reject","confidential_remarks":"The core exactness claim fails in Lemma 1 because the conditional Rician law depends on the phase of h_1 when A>0. This is not a cosmetic issue: the exact OP and ER expressions are the paper's main advertised contributions, and the subsequent bounds and UPA results all inherit the approximation. If the authors resubmit a version that reframes the analysis as an approximate method with quantified error and validates it against the phase-averaged exact law, the work could be reconsidered. The manuscript otherwise fits the journal's scope and contains a systematic treatment of ULA/UPA configurations."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colleague, here's my read of arXiv:2505.15200. The Rician FAS extension is a sensible next step for the fluid-antenna literature, and the UPA comparison is a nice addition. But the headline 'exact' results are not exact. The trouble starts in Appendix A: Eq. (A.1) conditions on x0,y0 and then treats the Rician noncentrality as if it only depends on |h1|. Under the paper's own model (Eq. 10), h_n given h1 has mean rho_n h1 + (1-rho_n)A; conditioning on |h1| leaves the phase of h1 in the noncentrality parameter, and that phase is not uniform. The cross term 2 rho sigma A x0 is dropped, and Eq. (A.2) also uses (1-rho_n^2)A^2 where the true expression would involve (1-rho_n)^2 A^2 and the cross term. So Lemma 1 holds exactly only for kappa=0. Theorem 1, the bounds, the ergodic-rate results, and the UPA formulas all inherit that approximation. The Monte Carlo agreement in Figs. 3-5 doesn't settle it, because the analytical curve is drawn from the same formula; a direct simulation of Eq. (10) is the right check.\n\nA second soft spot: Proposition 1 derives an infinite slope in N by holding |rho| fixed as N goes to infinity. But Eq. (9) makes rho_n a function of N, and for small n, rho_n -> 1 as N grows. So the argument doesn't support the claimed arbitrarily small outage with port count. The diversity-order claim in Proposition 2 is also bolted onto the same upper bound.\n\nWhat's genuinely useful: the kappa=0 limit reduces to known Rayleigh results, the UPA/ULA comparison is new, and the MRC benchmarking is a sensible performance reference. The paper is clearly written and the derivation structure is conventional. The free constant c>1 in Corollary 2 is an unspecified loose end but minor.\n\nWho should read this: researchers working on FAS performance analysis who want a Rician treatment. It deserves a serious referee; the flaw is subtle enough that a good reviewer could catch it and ask for a corrected derivation or an honest re-labeling as an approximation. But in current form I would not cite the exact formulas, and I'd want the phase-averaging fixed before it goes further.","headline":"The Rician FAS analysis is a useful extension, but the 'exact' joint PDF drops a phase-dependent cross term, and the high-N slope argument holds the wrong quantity fixed.","tokens_in":21752,"tokens_out":6417,"would_cite":false,"duration_ms":48780,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["94A05"],"pacs":[],"model":"deepseek-v4-flash","headline":"The paper claims that a fluid antenna receiver with N switchable ports under spatially correlated Rician fading has an outage probability that falls roughly with port count, with planar port layouts beating linear ones.","keywords":["fluid antenna system","Rician fading","outage probability","ergodic rate","spatial correlation","diversity order","uniform linear array","uniform planar array"],"falsifier":"Run a Monte Carlo simulation of the channel model in Eq. (10) with κ > 0, fix |h_1| = m_1, and compare the empirical distribution of |h_2| with the Rician density having noncentrality $\\sqrt$($ρ_2^{2}$ $m_1^{2}$ + (1−$ρ_2^{2}$)$A^{2}$) and variance $σ^{2}$(1−$ρ_2^{2}$); if the empirical distribution shifts with the phase of h_1, the joint PDF in Eq. (12) and the outage expression (15) are not exact. The same comparison at κ = 0 should match, isolating the phase-dependence error.","tokens_in":20626,"feed_emoji":"📡","tokens_out":5951,"duration_ms":49284,"temperature":0.7,"pith_summary":"This paper analyzes a receiver whose single antenna can be reconfigured among N positions, or ports, over a small space—a fluid antenna system—under Rician fading with spatial correlation between ports. It claims to provide exact expressions for the joint statistics of the port gains and for the outage probability, plus closed-form upper and lower bounds and ergodic-rate bounds. The central message is that the outage probability falls steeply as N grows, with diversity order approximately N, and that arranging ports on a planar grid (UPA) outperforms a linear arrangement (ULA) for the same number of ports. A sympathetic reading takes this as evidence that fluid antenna systems are a viable single-RF-chain alternative to conventional antenna arrays in line-of-sight-rich 6G environments.","feed_headline":"Fluid antennas: more ports steeper outage drop in Rician fading","feed_subtitle":"Closed-form analysis shows outage falls roughly with port count N, and planar layouts beat linear ones.","key_machinery":"The load-bearing object is a conditional-Rician joint distribution of the port amplitudes. It asserts that, with port 1 as reference, |h_n| given |h_1| = m_1 is Rician with noncentrality parameter $\\sqrt$($ρ_n^{2}$ $m_1^{2}$ + (1−$ρ_n^{2}$)$A^{2}$), where ρ_n is the J0 Bessel correlation between ports and A is the line-of-sight amplitude. This factorization reduces the N-dimensional outage integral to a single integral over m_1 with a product of Marcum Q-functions for the other ports, and it is what every later bound and diversity-order result inherits.","core_discovery":"The paper's central claim is that, under spatially correlated Rician fading, the port channel amplitudes of an Rx-SISO-FAS have a joint density of the product form in Eq. (12), in which each non-reference port's amplitude, conditioned on the reference port's amplitude, is Rician with a noncentrality parameter involving the reference amplitude and the line-of-sight component. From this joint density, the outage probability reduces to a one-dimensional integral (Theorem 1, Eq. (15)) whose integrand is the Rician density of the reference port times products of first-order Marcum Q-functions. The same machinery yields lower and upper outage bounds, closed-form ergodic-rate bounds, and asymptotic results: diversity order approximately N at high SNR, an infinite slope in N, and a zero slope in the Rician factor at large κ. The numerical results support the qualitative conclusions that FAS outperforms a fixed-position antenna, can beat an L-branch MRC system, and works better with UPA than ULA ports.","pith_inferences":["Editorial: Because the conditional-Rician step (A.2) ignores the phase of the reference channel, the formulas labeled exact should be read as approximations for κ > 0; the Monte Carlo agreement shown in the paper may hide errors at moderate κ or when ρ_n is large.","Editorial: The diversity-order claim d ≈ N likely survives the phase issue, since the N-fold product structure remains, but a proof that starts from the true conditional distribution could tighten or correct the order at finite SNR.","Editorial: The UPA-versus-ULA comparison suggests an optimization problem the paper does not solve: for a given rectangular footprint, the number of rows and columns that minimizes outage for fixed N can be chosen using the same correlation formulas.","Editorial: The same joint-distribution method could be extended to transmit-side FAS and dual-MIMO-FAS by conditioning on two reference ports, where the phase-dependence issue would need to be handled explicitly."],"forward_implications":["At high SNR, the outage probability behaves roughly like (1/γ_th)^N, so each additional port adds about one order of diversity; even a small fluid antenna with dozens of ports can reach very low outage.","For a fixed physical size, increasing the number of ports N lowers outage substantially, and the slope of the outage curve steepens with N, so port density matters more than array footprint.","A planar (UPA) port arrangement yields lower outage than a linear (ULA) one with the same N because the Euclidean port distances are larger and the Bessel correlations are smaller.","A single-RF-chain FAS can outperform an L-branch MRC combiner once N is sufficiently large (for example, N > 30 beats L = 5 in the paper's settings), at lower hardware cost.","A strong line-of-sight component helps at high SNR but degrades outage at low SNR, and increasing the Rician factor flattens the outage curve, so LoS strength should be treated as a system parameter rather than a universal blessing."],"supporting_citations":[{"why":"Introduces the performance limits of fluid antenna systems and the J0 spatial correlation model used in Eq. (9).","marker":"[40]"},{"why":"Establishes the fluid antenna system concept and the Rx-SISO-FAS channel model that this paper extends to Rician fading.","marker":"[41]"},{"why":"Supplies a spatial-correlation modeling approach for fluid antenna channels that the paper builds on.","marker":"[43]"},{"why":"Provides prior outage and diversity analysis for FAS under Rayleigh fading, which this paper generalizes to Rician fading.","marker":"[45]"},{"why":"Gives an earlier Nakagami-m FAS outage result in open form, motivating the closed-form Rician treatment here.","marker":"[47]"},{"why":"Defines the first-order Marcum Q-function used throughout the outage and ergodic-rate expressions.","marker":"[51]"},{"why":"Provides the Marcum Q-function approximation used to derive the upper outage bound.","marker":"[52]"},{"why":"Gives the MRC outage benchmark in Eq. (45) used for the FAS-versus-MRC comparison.","marker":"[53]"}],"fun_headline_variants":["Fluid antennas: adding ports slashes outage under Rician fading","Planar port layouts beat linear for fluid antennas in Rician fading","Fluid antenna analysis: N ports give diversity N, UPA beats ULA","Fluid antennas beat MRC under Rician fading with enough ports","More ports, less outage: fluid antennas in Rician fading"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The derivation of the joint density assumes that, once the magnitude of the reference port's channel is fixed, each other port's channel magnitude follows a Rician distribution whose noncentrality depends only on that magnitude—not on the phase of the reference channel; in the paper's own channel model the phase matters whenever a line-of-sight component is present, so the 'exact' expressions are unstated approximations for κ > 0 and exact only for κ = 0.","fun_headline_variants_meta":{"raw":{"variants":["Fluid antennas: adding ports slashes outage under Rician fading","Planar port layouts beat linear for fluid antennas in Rician fading","Fluid antenna analysis: N ports give diversity N, UPA beats ULA","Fluid antennas beat MRC under Rician fading with enough ports","More ports, less outage: fluid antennas in Rician fading"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001371,"raw_usage":{"total_tokens":5613,"prompt_tokens":1058,"completion_tokens":4555,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":674,"completion_tokens_details":{"reasoning_tokens":4462}},"tokens_in":674,"tokens_out":4555,"duration_ms":26373,"temperature":1.0,"reasoning_tokens":4462,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-07T15:23:49.571067+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Run a Monte Carlo simulation of the channel model in Eq. (10) with κ > 0, fix |h_1| = m_1, and compare the empirical distribution of |h_2| with the Rician density having noncentrality $\\sqrt$($ρ_2^{2}$ $m_1^{2}$ + (1−$ρ_2^{2}$)$A^{2}$) and variance $σ^{2}$(1−$ρ_2^{2}$); if the empirical distribution shifts with the phase of h_1, the joint PDF in Eq. (12) and the outage expression (15) are not exact. The same comparison at κ = 0 should match, isolating the phase-dependence error.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Defines the first-order Marcum Q-function used throughout the outage and ergodic-rate expressions."},{"cited_title":"Fluid antenna system: New insights on outage probability and dive rsity gain,","cited_arxiv_id":null,"evidence_quote":"Provides prior outage and diversity analysis for FAS under Rayleigh fading, which this paper generalizes to Rician fading."},{"cited_title":"Enhancing QoS through fluid antenna systems over corre lated Nakagami-m fading channels,","cited_arxiv_id":null,"evidence_quote":"Gives an earlier Nakagami-m FAS outage result in open form, motivating the closed-form Rician treatment here."},{"cited_title":"Asymptotic and numerical aspects of the nonc entral Chi- Square distribution,","cited_arxiv_id":null,"evidence_quote":"Provides the Marcum Q-function approximation used to derive the upper outage bound."},{"cited_title":"Ave rage outage duration of diversity systems over generalized fadi ng channels,","cited_arxiv_id":null,"evidence_quote":"Gives the MRC outage benchmark in Eq. (45) used for the FAS-versus-MRC comparison."}],"review_version":1}