{"id":"cf77b602-3a79-40f3-a1eb-0313c8d68559","arxiv_id":"2506.06754","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":4.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":1,"one_line_summary":"A joint beamforming and pinching-antenna position optimization for MIMO SWIPT is proposed, using WMMSE inner iterations and grid-search position updates.","lead":"This paper proposes a wireless system that uses movable pinching antennas on waveguides to send both data and power to separate receivers. Simulations suggest the approach can improve data rates while meeting energy harvesting targets, but the comparison uses fewer antennas in the baseline and the position search may miss the true optimum.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"At fc=28 GHz the outer-loop grid step (15 mm) exceeds half a wavelength (~5.4 mm), so the position update in P3 cannot resolve carrier-phase oscillations and the claimed joint position optimization and convergence are unsupported.","rationale":"The paper's core algorithmic claim is that jointly optimizing W and L with the AO framework maximizes sum-rate; the only mechanism for optimizing L is the grid search. Because the grid step is larger than the carrier-period phase scale, the argument that (P3) is 'effectively solved' fails, and the convergence guarantee, which relies on exact or at least monotone maximization of the surrogate, is not established. This is an internal inconsistency rather than a disagreement with the PASS channel model: even granting Eq. (3), the numerical procedure does not optimize the positions it claims to optimize. The reader's weakest-assumption pick (equal-amplitude coupling / no waveguide loss) is a physical modeling question shared with the prior PASS literature; it may be worth a sensitivity study, but it does not need to be true for the central algorithm to be correct within the paper's model. The unfair-baseline and lack-of-averaging points in the reader's rationale are also real, and a fair baseline and Monte Carlo averaging should be added, but the grid-resolution defect is the most load-bearing because it undermines the proposed optimization itself. Even so, the underlying idea is plausible and the derivation of the surrogate appears mostly sound, so a conditional accept with mandatory numerical revision is appropriate rather than rejection.","tokens_in":8433,"tokens_out":11593,"duration_ms":129058,"concrete_test":"Re-run Algorithm 1 on the same user drop and parameters (Pmax=43 dBm, Fig. 3 setup) with the outer-loop grid refined to L=20001 (1.5-mm step) or, better, with a continuous refinement step (e.g., golden-section or gradient ascent on the surrogate in (P3) around the best grid point). Record final sum-rate, the selected PA positions, and the convergence curve. If the sum-rate changes by more than a small amount or the positions move by more than a wavelength, the published curves are not the result of the claimed joint optimization and the central claim needs revision.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Load-bearing concern: Algorithm 1's outer loop solves (P3) by a one-dimensional grid search with L=2001 points over Lx=30 m, giving a step of 15 mm. With fc=28 GHz, lambda=10.7 mm, and n_ref=1.44, the channel phase in Eq. (3) contains exp(-j kappa(D + n_ref l)); adjacent grid points advance this phase by roughly 1.4 lambda of distance, i.e., multiple full cycles. The objective in (P3) is therefore highly oscillatory in l, and the grid samples it far below the Nyquist rate. The selected grid point is essentially arbitrary with respect to carrier phase, so the Gauss-Seidel update is not an effective maximizer of (P3), and the claimed monotone non-decrease of the objective no longer follows. Consequently Figs. 2-4 do not demonstrate the jointly optimized PA positions that the paper's central claim requires; the curves can change materially under a finer grid or a continuous local search. This is a correctness risk in the paper's own optimization procedure, independent of whether the equal-amplitude channel model in Eq. (3) is realistic. The reader's channel-model concern is reasonable but secondary; the grid-resolution issue is the place where the argument is least secure.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies a downlink MIMO pinching-antenna system (PASS) for simultaneous wireless information and power transfer (SWIPT). It formulates a joint optimization of the transmit beamforming matrices and the pinching-antenna positions to maximize the sum-rate of multiple information-decoding receivers subject to per-receiver harvested-energy, transmit-power, and antenna-spacing constraints. The proposed solution alternates between a WMMSE-based inner loop for the beamformers and a Gauss-Seidel outer loop in which each pinching-antenna position is updated by a one-dimensional grid search. Numerical results compare the proposed design against conventional MIMO, fixed-position PASS, and zero-forcing benchmarks and report substantial gains.","tokens_in":8640,"tokens_out":7408,"duration_ms":77410,"significance":"If the numerical claims are reliable, the paper makes a useful contribution: it demonstrates that movable pinching antennas can provide additional spatial degrees of freedom for SWIPT, jointly serving information and energy receivers. The analytical framework is standard—the WMMSE surrogate and the successive-convex-approximation treatment of the energy constraints are correctly applied—and the benchmark comparisons are well chosen. The derivation is self-contained and introduces no fitted constants. The main weakness is the numerical position-optimization step, whose grid resolution is too coarse to resolve the carrier-phase structure of the channel, and the paper does not provide evidence that the reported curves are stable with respect to that resolution.","major_comments":[{"comment":"The outer-loop position update in Algorithm 1 is not reliable because the grid in (P3) is too coarse to resolve the carrier phase. With Lx=30 m and L=2001, the step is 15 mm. At fc=28 GHz and n_ref=1.44, the phase term κ n_ref l in Eqs. (3) and (24) advances by roughly 2π × 1.44 × 0.015 / 0.0107 ≈ 4.0π rad between adjacent grid points, i.e., more than one full period; the distance-dependent phase κD(l) varies as well. The objective of (P3) is therefore highly oscillatory in lm,n, and the grid samples it below the Nyquist rate. The selected grid point is effectively arbitrary with respect to carrier phase, so the Gauss-Seidel update is not a reliable maximizer of (P3), the monotone-convergence argument in Section III-3 does not establish convergence to a meaningful solution of (P1), and the position-optimization gains shown in Figures 2–4 are not substantiated. Please rerun with a step at most λ/(2 n_ref) ≈ 3.7 mm (e.g., L ≥ 8000) or with a continuous local refinement, and show that the conclusions are stable in L.","section":"III-2, III-3, IV"},{"comment":"The numerical section does not report whether the energy-harvesting constraints are actually feasible after the position update. The coarse grid may skip feasible positions satisfying (28), and the paper does not describe how an infeasible grid point is handled in Algorithm 1. The authors should report the achieved minimum harvested energy across EHRs and the feasibility rate for each benchmark, since the reported sum-rate improvement could otherwise come from violating constraint (11a).","section":"IV"}],"minor_comments":[{"comment":"Constraint (29a) contains a typo: it should read lm,n′ − lm,n ≥ L0 for the adjacent pinching antennas, but as printed, lm,n′ − lm,n′ is identically zero.","section":"III-2"},{"comment":"The symbol L is used both for the position matrix in the system model and for the number of grid-search points in the sentence defining Gm; please rename one of them to avoid ambiguity.","section":"III-2"},{"comment":"In the system model, the phrase 'while that of EHR k is located' should refer to EHR q, not k.","section":"II"},{"comment":"The stated complexity O(M^3 J^3 K^3) covers only the inner QCQP solve; the outer loop additionally performs M N one-dimensional grid searches of size L. The total complexity should be stated accordingly.","section":"III-3"},{"comment":"The equal-amplitude, lossless-waveguide channel model in Eq. (3) is an idealization adopted from [9]. The paper should state explicitly that unequal coupling between pinching elements or non-negligible waveguide loss would reduce the coherent beamforming gain, and ideally include a sensitivity check.","section":"II, IV"},{"comment":"No code or data are provided. Reporting the dependence of the results on the grid resolution L (e.g., L=2001 versus L=8000) would substantially increase confidence in the numerical claims.","section":"IV"}],"recommendation":"major_revision","confidential_remarks":"The grid-resolution issue is the main barrier. It is local and fixable: rerunning the position search with a smaller step or a continuous local refinement, and reporting the corresponding curves, would likely settle the matter. I would also ask the authors to report EH feasibility and the number of grid points used in the final figures, because the current manuscript does not allow a reader to verify that the reported gains are not artifacts of the undersampled position search."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colleague,\n\nThe paper is the first to bring pinching-antenna systems (PASS) to SWIPT, and that part is legitimate: the system model, the WMMSE inner loop, and the Gauss-Seidel position update are derived carefully, and the benchmarks include fixed-PA and ZF designs. If the numerical results held up, this would be a useful extension of PASS to energy-constrained IoT.\n\nThe numerical results do not hold up. The outer loop solves (P3) by a grid search with L=2001 points over a 30 m aperture, i.e. a 15 mm step. At 28 GHz with n_ref=1.44, the channel phase in Eq. (3) advances by roughly 845 rad/m in l, so the grid steps over two full cycles of the carrier phase. The objective in (P3) is highly oscillatory in l, and a 15 mm grid samples it far below the Nyquist rate. The chosen PA position is therefore effectively arbitrary with respect to the carrier phase; the Gauss-Seidel update is not an effective maximizer, and the claimed monotone convergence of Algorithm 1 no longer follows. Consequently Figs. 2–4 do not demonstrate the jointly optimized positions that the central claim requires. This is not a hardware-modeling quibble; it is a flaw in the paper's own optimization procedure.\n\nTwo more issues are secondary but worth naming. The \"conventional MIMO\" benchmark uses M=4 antennas while the PASS uses 4 waveguides × 3 PAs = 12 radiating elements; that comparison loads the dice. And the channel model in Eq. (3), borrowed from [9], assumes equal coupling and no waveguide loss; real hardware may diverge, though that is a modeling simplification rather than a mistake in the derivation. The paper also reports no averaging over random user drops, so the plotted curves may not be typical.\n\nThe conceptual contribution—a PASS-aided SWIPT formulation with WMMSE-based beamforming—is sound and worth publishing once the position-optimization loop is fixed, e.g., by a fine grid (step <= lambda_g/4), a local continuous search, or a gradient step. I would send this to peer review, but I would require the numerical section to be redone before acceptance. The paper is not ready as-is.","headline":"First PASS–SWIPT formulation with a plausible WMMSE framework, but the coarse position grid search at 28 GHz undermines the numerical claims and the optimization's convergence guarantee.","tokens_in":9217,"tokens_out":3249,"would_cite":false,"duration_ms":35714,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["94A15","90C26","78A50"],"pacs":[],"model":"deepseek-v4-flash","headline":"Jointly moving pinching antennas and optimizing transmit beams maximizes SWIPT sum-rate while meeting per-receiver energy targets; simulations show gains over conventional MIMO.","keywords":["pinching antenna systems","simultaneous wireless information and power transfer","MIMO beamforming","wireless power transfer","antenna position optimization","WMMSE","alternating optimization","dielectric waveguides"],"falsifier":"Place one waveguide with $N$ tunable pinching antennas in an anechoic chamber, fix all but one PA, and measure the received signal as the free PA slides along the guide; if the measured amplitude varies by more than a few dB with position, or if the phase does not follow $\\exp\\{-\\mathrm{j}\\kappa(D+n_{\\mathrm{ref}}l)\\}$, then Eq. (3)'s equal-amplitude coherent-sum model is falsified and the predicted sum-rate gains need re-evaluation.","tokens_in":8203,"feed_emoji":"📡","tokens_out":6849,"duration_ms":64731,"temperature":0.7,"pith_summary":"This paper introduces pinching-antenna systems (PASS) into simultaneous wireless information and power transfer (SWIPT): an access point feeds several dielectric waveguides, each carrying multiple movable pinching antennas, to serve information-decoding receivers and energy-harvesting receivers at the same time. It asks how to place the antennas and choose the transmit beamforming matrices so that the total information rate is as large as possible while every energy harvester still collects at least a required power. The paper answers with an alternating optimization algorithm, alternating a WMMSE-based beamforming update with a Gauss-Seidel position update solved by one-dimensional grid search. If the reported numbers hold, movable pinching antennas give SWIPT a practical way to obtain spatial degrees of freedom that fixed-array MIMO cannot reach.","feed_headline":"Moving pinching antennas boosts SWIPT sum-rate past fixed-array MIMO","feed_subtitle":"Jointly positioning antennas and shaping beams lets one 28 GHz link deliver data and guaranteed energy to IoT users.","key_machinery":"The load-bearing object is the effective channel coefficient in Eq. (3), where each waveguide's $N$ pinching antennas contribute a coherent sum of equal-amplitude, phase-shifted paths: $h_{m,k,j}(\\mathbf{l}_m) = \\xi \\sum_{n=1}^N \\frac{\\exp\\{-\\mathrm{j}\\kappa(D_{m,k,j}(l_{m,n})+n_{\\mathrm{ref}}l_{m,n})\\}}{\\sqrt{N}\\,D_{m,k,j}(l_{m,n})}$. A pinching antenna is a radiator that taps the guided wave on a dielectric waveguide; sliding it changes both the free-space distance $D$ and the guided-wave phase $n_{\\mathrm{ref}}l$. Around this channel, the paper builds a concave WMMSE surrogate for the rate, an SCA lower bound on harvested energy, and a Gauss-Seidel loop that reduces position optimization to scalar one-dimensional searches over per-PA positions. This decomposition turns a highly non-convex, NP-hard problem into a solvable alternating sequence of convex QCQPs and scalar grid searches.","core_discovery":"The central claim is that jointly optimizing the pinching beamforming matrices $\\mathbf{W}$ and the pinching-antenna positions $\\mathbf{L}$ in a MIMO PASS-aided downlink maximizes the sum-rate of all information-decoding receivers subject to a minimum harvested energy at each energy-harvesting receiver, and that this joint design outperforms conventional MIMO, fixed-PA, and ZF-based benchmarks. The argument is algorithmic: the non-convex sum-rate objective is replaced by a concave WMMSE surrogate, the energy constraints are replaced by successive-convex-approximation lower bounds, and each PA position is updated by solving a scalar grid-search problem while holding everything else fixed. Because every update is non-decreasing and the feasible sets are bounded, the paper argues the scheme converges. The numerical section reports that the gains persist across transmit-power and antenna-count sweeps, with convergence within about 20 iterations.","pith_inferences":["An immediate extension the paper leaves implicit is to treat each pinching antenna's coupling amplitude as an additional optimization variable; if real waveguides couple unequally, optimizing amplitudes alongside positions would likely recover most of the computed gains.","The same WMMSE-plus-Gauss-Seidel scaffold transfers to other PASS objectives, such as max-min harvested energy or secrecy-rate maximization, because the position update is already reduced to scalar searches that do not depend on the specific objective.","The near-field, LoS-strengthening argument suggests a testable prediction: the performance gap over fixed arrays should widen as receivers move closer to the waveguides, where distance control matters most.","A hardware or ray-tracing check of Eq. (3)'s equal-amplitude coherent-sum assumption would tell whether the 28 GHz results transfer to practical waveguides with loss."],"forward_implications":["If the central claim is right, PASS-aided SWIPT can serve IoT devices' data and charging needs from one aperture, with higher sum-rate than conventional MIMO at the same transmit power.","Position optimization, not just beamforming, is what delivers the gain: the fixed-PA benchmark falls far below the joint design in the paper's Fig. 3.","Adding more pinching antennas per waveguide steadily raises sum-rate, implying the architecture scales by adding low-cost radiators on existing waveguides.","A ZF-based position design approaches the proposed scheme at high transmit power, so simpler beamforming may suffice in high-SNR regimes.","The algorithm's convergence within about 20 iterations and its per-iteration complexity $O(M^3J^3K^3)$ make it a tractable candidate for real-time position and beam updates."],"supporting_citations":[{"why":"Supplies the equal-amplitude coherent-sum channel model in Eq. (3), including the $1/\\sqrt{N}$ attenuation and negligible-waveguide-loss assumption on which the optimization rests.","marker":"[9]"},{"why":"Provides the WMMSE-to-rate equivalence used to build the concave surrogate function for the sum-rate objective.","marker":"[13]"},{"why":"Contributes the lemma showing the surrogate is concave and upper-bounds the achievable rate, which justifies the inner-loop transformation.","marker":"[14]"},{"why":"Supplies the linear energy-harvesting model $E_q = \\eta_q \\mathrm{Tr}(\\cdot)$ that defines the per-EHR constraints.","marker":"[4]"},{"why":"Frames pinching antennas as a flexible-antenna technology and motivates treating PA positions as design variables.","marker":"[7]"},{"why":"Motivates the PASS architecture of multiple waveguides with movable pinching antennas that the system model adopts.","marker":"[8]"}],"fun_headline_variants":["Joint PA positioning and beamforming boosts SWIPT sum-rate","Moving pinching antennas improves SWIPT data and energy delivery","Jointly optimizing PA positions and beams boosts SWIPT","Pinching-antenna placement and beamforming lift SWIPT rates"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The central premise is that every pinching antenna on a waveguide radiates the fed signal with the same amplitude $1/\\sqrt{N}$ and that waveguide propagation loss is negligible, so the total channel is a clean coherent sum of equal-amplitude paths; if real hardware couples unevenly or loses energy along the guide, the reported beamforming and positioning gains may shrink.","fun_headline_variants_meta":{"raw":{"variants":["Joint PA positioning and beamforming boosts SWIPT sum-rate","Moving pinching antennas improves SWIPT data and energy delivery","Jointly optimizing PA positions and beams boosts SWIPT","Pinching-antenna placement and beamforming lift SWIPT rates"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000476,"raw_usage":{"total_tokens":2352,"prompt_tokens":925,"completion_tokens":1427,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":541,"completion_tokens_details":{"reasoning_tokens":1359}},"tokens_in":541,"tokens_out":1427,"duration_ms":10073,"temperature":1.0,"reasoning_tokens":1359,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-07T05:50:04.012951+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Place one waveguide with $N$ tunable pinching antennas in an anechoic chamber, fix all but one PA, and measure the received signal as the free PA slides along the guide; if the measured amplitude varies by more than a few dB with position, or if the phase does not follow $\\exp\\{-\\mathrm{j}\\kappa(D+n_{\\mathrm{ref}}l)\\}$, then Eq. (3)'s equal-amplitude coherent-sum model is falsified and the predicted sum-rate gains need re-evaluation.","supporting_citations":[{"cited_title":"MIMO –PASS: Uplink and downlink transmission via MIMO pinching-antenn a systems","cited_arxiv_id":null,"evidence_quote":"Supplies the equal-amplitude coherent-sum channel model in Eq. (3), including the $1/\\sqrt{N}$ attenuation and negligible-waveguide-loss assumption on which the optimization rests."},{"cited_title":"Joint transmit b eamforming and RRH selection for user–centric green MIMO C–RAN,","cited_arxiv_id":null,"evidence_quote":"Contributes the lemma showing the surrogate is concave and upper-bounds the achievable rate, which justifies the inner-loop transformation."},{"cited_title":"Energy efficiency optimization with SWIPT in MIMO broadcast channe ls for internet of things,","cited_arxiv_id":null,"evidence_quote":"Supplies the linear energy-harvesting model $E_q = \\eta_q \\mathrm{Tr}(\\cdot)$ that defines the per-EHR constraints."},{"cited_title":"Flexible-ante nna systems: A pinching–antenna perspective,","cited_arxiv_id":null,"evidence_quote":"Frames pinching antennas as a flexible-antenna technology and motivates treating PA positions as design variables."},{"cited_title":"Pinch ing an- tenna systems (PASS): Architecture designs, opportunitie s, and outlook,","cited_arxiv_id":null,"evidence_quote":"Motivates the PASS architecture of multiple waveguides with movable pinching antennas that the system model adopts."}],"review_version":1}