{"id":"5e724c35-cabd-4bde-849f-545fd8250abc","arxiv_id":"2505.17526","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":3,"one_line_summary":"Beamformer switching between two weight vectors modulates a secondary user's observed interference into a low-rate OOK stream, with summed spectral efficiency gains only for low channel correlation.","lead":"A transmitter can serve a primary user while sending a second, low-rate data stream to an extra user by switching between two beamforming settings that change how much interference that user sees. This paper derives the error rate and sum-rate trade-off for this interference modulation scheme, aimed at low-power IoT devices in 6G networks.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Perfect-CSI assumption for the SU null is the load-bearing risk: with channel estimation error the OOK '0' leaks signal, raising the error floor and eroding the claimed sum-rate gain.","rationale":"The reader's weakest_assumption identifies the same load-bearing concern: perfect CSI for the SU is required to make the OOK '0' state truly silent, and any estimation error leaks residual energy that raises the error floor. I agree that this is the most consequential weakness, because the central claim of a sum-rate increase depends on Pe < 10^-5 being achievable at small alpha. I examined the algebraic issues the reader raised. The factor-of-2 error in Eq. (5) makes the computed xi larger than the correct value, so the paper's efficiency numbers are conservative rather than optimistic; correcting it strengthens the low-correlation claim. The possible extra |omega_1|^2 factor in Eq. (8) is similarly conservative in the low-correlation regime. Thus these algebra errors do not invalidate the central claim, though they should still be corrected. The missing baseline comparison to conventional multiple access is a presentation issue, not a correctness issue, because the claim as stated is relative to PU-only operation. The perfect-CSI assumption, however, is structural: without it, OOK zero is not zero, and the BER analysis in Eqs. (12)-(15) no longer applies. Therefore the verdict remains conditional, as the reader concluded, pending a demonstration that practical CSI accuracy is sufficient or a robustness analysis showing the scheme tolerates realistic estimation error.","tokens_in":7467,"tokens_out":36864,"duration_ms":308375,"concrete_test":"Simulate with estimated SU channel h_hat_SU = h_SU + e, e ~ CN(0, sigma_e^2 I), and compute the beamforming weights from h_hat_SU instead of h_SU. Evaluate the SU bit-error probability using Eq. (15) with the bit-0 distribution replaced by Gamma(N, sigma_n^2 + beta), where beta is the residual signal power |h_SU^T omega_0|^2. Sweep sigma_e^2 for alpha = 0.1, gamma = 30 dB, rho = 0.2, and record the largest sigma_e^2 (as a fraction of the channel gain) that still meets Pe < 10^-5. If that tolerance is below what TDD reciprocity or limited feedback can provide, the practical claim needs an explicit CSI-acquisition condition.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central sum-rate claim rests on Eq. (5) being exactly implementable: the transmitter must design omega_0 with h_SU^T omega_0 = 0 and omega_1 with h_SU^T omega_1 = sqrt(alpha). The OOK '0' symbol is therefore silent at the SU only if the SU channel is known perfectly at the instant of transmission. In a mobile or time-varying setting, the channel used for weight design differs from the actual channel, so the residual signal at the SU in bit 0 is no longer zero. The bit-0 energy distribution in Eq. (12) then becomes a Gamma(N, sigma_n^2 + residual) mixture instead of Gamma(N, sigma_n^2), and the error probability in Eq. (15) acquires an irreducible floor that cannot be removed by increasing N. For small alpha, the regime advertised for 'minimal power allocation', the bit-1 SNR is low, so even a modest CSI error can make Pe < 10^-5 unattainable. The paper provides no CSI acquisition protocol, no sensitivity analysis, and no characterization of the maximum allowable estimation error. The factor-of-2 algebra error in Eq. (5)/Eq. (6) is real but conservative, so it does not threaten the claim; the CSI assumption does, because the entire OOK construction fails without it.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper proposes 'interference modulation,' a downlink technique in which a single-RF-chain multi-antenna transmitter serves a primary OFDM user and simultaneously sends low-rate on-off-keying data to a secondary user by switching between two beamforming vectors. The design constrains the two beamformers so that the PU receives constant power (1−α) while the SU receives either a controlled interference level (bit 1) or a null (bit 0). The paper derives the beamformer norms (Eq. (5)), an average power-normalization factor ξ (Eq. (6)), a closed-form OOK bit-error probability (Eq. (15)), and a sum-rate expression (Eq. (16)). Monte Carlo BER simulations are reported to match the analytical curves, and sum-rate plots are used to claim spectral-efficiency gains for channels with similar gains and low correlation.","tokens_in":7742,"tokens_out":12664,"duration_ms":94747,"significance":"The conceptual contribution is genuine and timely: it adds a low-rate user channel without an extra RF chain, extra time slot, or independent superimposed waveform, by exploiting the spatial degrees of freedom already used for the PU. The analytical BER derivation is self-contained and is confirmed by simulation in Fig. 5; the selection of N_alpha from Eq. (15) is an operational rule rather than a fit to data. The main claims, however, rest on two assumptions that need to be made explicit and tested: exact instantaneous CSI at the transmitter for the SU null, and the corrected algebra in the power-normalization factor. If these are addressed, the paper would be a useful contribution to low-rate/low-power multiple access for 6G RedCap scenarios.","major_comments":[{"comment":"The expression for |ω1|^2 drops the factor 2 in the cross term. Substituting b_SU = √α e^{−j arg ρ} and b_PU = √(1−α) into Eq. (4) gives |ω1|^2 = (1 − 2|ρ|√(α(1−α)))/(1 − |ρ|^2), while Eq. (5) has a single factor. Consequently, Eq. (6) should read ξ = (2 − α − 2|ρ|√(α(1−α)))/(2(1 − |ρ|^2)). Since ξ enters the sum-rate expression in Eq. (16), Figs. 3 and 6 and all quantitative sum-rate conclusions need to be recomputed with the corrected ξ. The error is conservative in direction (the corrected ξ is smaller, so the published curves understate the power efficiency), but the numerical results as printed are not the ones the model predicts.","section":"Section III-A, Eqs. (5) and (6)"},{"comment":"The entire OOK '0' detection rests on h_SU^T ω0 = 0, which requires perfect instantaneous CSI of the SU channel at the transmitter. With any channel estimation error, the '0' symbol contains residual signal power, so Eq. (12) should be a Gamma mixture with an additional scale parameter, and the error probability in Eq. (15) acquires a floor that cannot be lowered by increasing N. The paper gives no CSI acquisition protocol, no model of estimation error, and no sensitivity analysis, even though the claim that the scheme works with 'minimal power allocation' depends directly on this assumption. Please add an error-robustness analysis or state explicitly that the reported gains are ideal upper bounds under perfect CSI.","section":"Section III-A and Section III-B, Eqs. (5), (12), (15)"},{"comment":"The normalization of ω0 and ω1 is not fully specified. ξ is defined as the average of the squared norms, while Eq. (16) treats ξ as an SNR attenuation factor; this is consistent only if both beamformers are scaled by 1/√ξ before transmission. The manuscript should state this scaling explicitly and give the normalized transmitted signal model. Otherwise a reader who divides the weights by ξ rather than by √ξ will obtain a different power budget and different sum-rate behavior.","section":"Section III-A, Eq. (6), and Section IV, Eq. (16)"}],"minor_comments":[{"comment":"The Gaussian PDF is written as f_S(s) = 1/(√(π σ^2)) e^{x^2/σ^2}; the variable x should be s, and for a real Gaussian with variance σ^2 the normalization should be 1/(√(2π)σ) with exponent −x^2/(2σ^2). Please correct the expression and state whether σ^2 denotes the variance of the real part, the imaginary part, or the complex sample.","section":"Section III-B, Eq. (7)"},{"comment":"The phrase 'Nα denotes the highest bandwidth ratio' is inconsistent with its use in Eq. (16), where Nα is the integration length in samples. Since the SU rate is 1/Nα, the quantity to maximize is the bandwidth ratio 1/Nα, so Nα should be described as the smallest number of samples (equivalently, the largest bandwidth ratio) that meets the Pe < 10^−5 target.","section":"Section IV, Eq. (16)"},{"comment":"The simulation setup is not fully specified: the figure caption and text do not state the OFDM parameters (number of subcarriers M, sample count N per OOK symbol), the value of α used for the BER curves, or how the beamformers in Eq. (5) were constructed in the Monte Carlo simulation. Please provide these details for reproducibility.","section":"Section IV, Fig. 5"},{"comment":"The sentence 'If the number of transmitter antennas is greater than 2' should read 'at least 2' (or 'greater than 1'): for K=2 the constraint system is exactly determined rather than underdetermined, and the same minimum-norm formula applies.","section":"Section III-A"},{"comment":"The numerical evaluation compares only against the PU-only baseline. Since the title and introduction position the work as a multiple-access technique, a comparison with at least one existing low-rate multiple-access scheme (e.g., NOMA, time-sharing, or beam index modulation) would help the reader assess the practical efficiency of the proposed method.","section":"Section IV and Section I"},{"comment":"The text says that in the OOK '0' state the signal is 'scattered,' but the constraint in Section III-A is a perfect null at the SU (h_SU^T ω0 = 0). Please align the wording with the mathematical model, since scattering usually implies nonzero residual energy.","section":"Section II"}],"recommendation":"major_revision","confidential_remarks":"The idea is novel enough for a communications journal, and the analytical BER derivation is a strength. However, the missing factor of two in Eqs. (5)–(6) affects all numerical efficiency claims, and the perfect-CSI assumption is not discussed despite being essential to the OOK construction. I would ask the authors to fix the algebra, rerun the figures, and add a robustness discussion before considering acceptance. The paper may also benefit from a more careful comparison with NOMA and index-modulation baselines to justify the 'multiple access' framing."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Here's my read on 2505.17526. The core idea is real: switch between two analog beamforming vectors so that a secondary user sees OOK-modulated interference power, while the primary user's OFDM is barely affected. That's a genuinely different way to piggyback a low-rate RedCap channel onto a single-RF-chain transmitter, and it doesn't need extra hardware. The paper explains it clearly, and the BER analysis is self-consistent; the simulated BER matches the energy-detection theory.\n\nThe soft spots are mostly in the quantitative claims. First, Eq. (5) and (6) have a factor-of-2 error in the cross term of |omega_1|^2. The text says b_SU = sqrt(alpha) e^{-j theta}, which gives 2 Re{rho b_SU b_PU*} = 2|rho| sqrt(alpha(1-alpha)), but Eq. (5) writes 1 - |rho| sqrt(alpha(1-alpha)) and Eq. (6) follows suit. The error is conservative—it makes the power loss look larger than it is—so the sum-rate curves are slightly pessimistic, not optimistic. Still, it needs fixing. There's also a phase inconsistency: the text uses e^{-j theta}, the vector in Eq. (5) writes e^{j arg(rho)}.\n\nThe bigger concern is the perfect-CSI assumption for the SU null. The entire bit-0 symbol depends on h_SU^T omega_0 = 0 exactly. With any estimation error, residual signal leaks into the '0' slot, and the OOK error probability gets an irreducible floor that no amount of averaging can remove. The paper doesn't provide a sensitivity analysis or a CSI acquisition protocol. In a mobile setting this could kill the advertised gains. This is a standard limitation in beamforming papers, but here it's structural—the secondary receiver is an energy detector, so it can't average out a constant residual.\n\nThird, the sum-rate evaluation compares only against serving the PU alone. That's a useful sanity check, but it's not a fair multiple-access baseline. A time-sharing or NOMA comparison would put the 'efficiency gain' in context. The authors should add that.\n\nOverall, the idea is plausible and the paper is honest about its low-rate scope. The flaws are fixable. It deserves a serious referee, but I wouldn't cite it until the normalization is corrected and the CSI sensitivity is addressed.","headline":"Real new idea for piggybacking a low-rate OOK channel onto a single-RF-chain beamformer, but the sum-rate claims need a factor-of-2 fix and a much deeper look at CSI sensitivity before I'd trust the numbers.","tokens_in":8264,"tokens_out":7519,"would_cite":false,"duration_ms":45544,"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":"A single-RF-chain transmitter can serve a second low-rate user by encoding OOK bits into the interference its primary OFDM signal already creates, switching between two beamforming weight vectors that keep the primary user's power near…","keywords":["interference modulation","on-off keying","analog beamforming","sum rate","channel correlation","single RF chain","6G RedCap","metasurface"],"falsifier":"Take a transmitter using the proposed weights, add a known mismatch between the channel used to design $\\omega_0$ and the actual secondary-user channel (for instance, a 5% norm error or a small angular drift), and measure the OOK bit-error rate as a function of $N$. If the error rate stops reaching $10^{-5}$ at the predicted $N_\\alpha$ even with the designed $\\alpha$, or if the measured sum rate drops below the primary-only rate, the central claim fails in that regime.","tokens_in":7281,"feed_emoji":"📶","tokens_out":7486,"duration_ms":54376,"temperature":0.7,"pith_summary":"The paper claims that a transmitter with one radio-frequency chain and multiple antennas can add a low-rate second user almost for free by encoding that user's bits in the interference its primary OFDM signal already creates at the secondary receiver. The transmitter alternates between two beamforming configurations: one focuses a fraction $\\alpha$ of the power toward the secondary user (an OOK '1'), and the other nulls the signal at that user entirely (an OOK '0'), while the primary user's received power stays near $1-\\alpha$. The authors derive closed-form weights, an efficiency measure $\\xi$, and a sum-rate expression, and they show that simulated bit-error rates match the theoretical OOK analysis. If the scheme works as described, 6G base stations and metasurface front ends could serve reduced-capability devices with minimal extra power and no new hardware.","feed_headline":"Interference becomes a second data link in a single-RF-chain system","feed_subtitle":"Switching beamforming weights encodes OOK bits in the primary signal's spill, adding RedCap user at minimal power cost.","key_machinery":"The central object is the pair of beamforming weight vectors $(\\omega_0, \\omega_1)$, computed via the Moore-Penrose pseudoinverse of the channel matrix $A = [h_{\\mathrm{SU}} \\, h_{\\mathrm{PU}}]$, together with the average normalization factor $\\xi$. $\\omega_0$ nulls the secondary user while delivering $\\sqrt{1-\\alpha}$ to the primary user; $\\omega_1$ delivers $\\sqrt{\\alpha}$ to the secondary user and $\\sqrt{1-\\alpha}$ to the primary user. Since the constraints are underdetermined for more than two antennas, minimum-norm solutions exist, and $\\xi$ is the average squared norm of the two weights, measuring the power cost of satisfying both users' constraints. The second mechanism is the OOK energy detector: the secondary receiver integrates $N$ samples of the OFDM waveform, and the threshold $\\delta^*$ is derived from the Gamma distributions of the 'on' and 'off' energies. These two pieces combine into the sum-rate formula.","core_discovery":"The paper's central claim is that beamforming weights can be designed so that a secondary user observes a controlled on-off energy pattern while the primary user's receiver sees almost no change. With unit-norm channel vectors $h_{\\mathrm{PU}}$ and $h_{\\mathrm{SU}}$ whose correlation is $\\rho = \\langle h_{\\mathrm{SU}}, h_{\\mathrm{PU}}\\rangle$, the transmitter chooses $\\omega_0$ and $\\omega_1$ such that $h_{\\mathrm{SU}}^T \\omega_0 = 0$, $h_{\\mathrm{SU}}^T \\omega_1 = \\sqrt{\\alpha}$, and $h_{\\mathrm{PU}}^T \\omega_0 = h_{\\mathrm{PU}}^T \\omega_1 = \\sqrt{1-\\alpha}$. The Moore-Penrose solution gives $|\\omega_0|^2 = (1-\\alpha)/(1-|\\rho|^2)$ and $|\\omega_1|^2 = (1 - \\sqrt{\\alpha}\\sqrt{1-\\alpha}|\\rho|)/(1-|\\rho|^2)$, normalized by the average factor $\\xi$ from Eq. (6). The secondary receiver integrates $N$ OFDM samples and applies a threshold derived from the Gamma distributions of signal-plus-noise and noise-only energies, yielding a closed-form OOK error probability. Sum rate is then $R = \\log_2(1 + \\gamma/(\\xi(1-\\alpha))) + 1/N_\\alpha$, with $N_\\alpha$ chosen so that $P_e < 10^{-5}$. The paper reports that for channels with similar gains and low correlation this sum rate exceeds the primary-only rate, and that simulated BER matches the theoretical curve.","pith_inferences":["A natural test the paper does not run is to add channel estimation noise to the design of $\\omega_0$ and re-measure the OOK error rate; if a few percent of channel error pushes $P_e$ above $10^{-5}$, the practical gain window will be much narrower than Fig. 6 suggests.","Because the OOK '1' symbol is simply the primary OFDM waveform focused toward the secondary user, the scheme only works while the primary user is actively transmitting; scheduling the secondary stream around primary silent periods would be needed in a real deployment.","The same interference-modulation idea could be applied to a reconfigurable intelligent surface acting as the transmitter front end, which the authors list as future work; an experiment with a metasurface would directly test hardware feasibility.","The sum-rate formula assumes a fixed $\\alpha$ and independently optimized $N_\\alpha$; jointly optimizing $\\alpha$ with the primary user's modulation order and the secondary user's target rate could shift the reported gains."],"forward_implications":["A single RF chain plus an analog beamformer can carry a second low-rate stream without additional hardware, raising sum rate when the two users' channels are weakly correlated and similar in gain.","The transmitter has a closed-form design rule: choose $\\alpha$ and the integration length $N$ to meet the secondary user's target bit-error probability, then read off the primary user's capacity loss from $\\xi$.","The secondary receiver is just an energy detector; it needs no channel estimation or interference cancellation.","The achievable secondary rate is capped at 1 bit/s/Hz by the OOK alphabet, so the technique targets low-rate reduced-capability links rather than high-throughput multiplexing.","The method extends to metasurface-based front ends, since the only requirement is the ability to switch between two radiation patterns."],"supporting_citations":[{"why":"Supplies the Moore-Penrose pseudoinverse method used to solve the underdetermined beamforming equations for $\\omega_0$ and $\\omega_1$.","marker":"[15]"},{"why":"Baseline single-RF-chain analog-beamforming beam index modulation that the proposed scheme extends and contrasts with.","marker":"[10]"},{"why":"Recent RIS-assisted beam index modulation that requires specially designed receivers, against which the no-extra-hardware claim is positioned.","marker":"[14]"},{"why":"Provides the index modulation concept of conveying extra information through the spatial domain that motivates embedding secondary-user data in beamformer settings.","marker":"[9]"}],"fun_headline_variants":["Beamforming encodes secondary data in interference","Interference shaping yields a minimal-power OOK channel","Turn spill into a second data link in one RF chain","Low-rate data via controlled interference in 6G","Modulating interference powers a RedCap side channel"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that the transmitter knows both users' channels exactly and can compute weight vectors that put exactly zero signal energy at the secondary user during an OOK '0'; if that estimate is imperfect or the channel changes, residual energy leaks into the '0' symbol and the predicted error rate no longer holds.","fun_headline_variants_meta":{"raw":{"variants":["Beamforming encodes secondary data in interference","Interference shaping yields a minimal-power OOK channel","Turn spill into a second data link in one RF chain","Low-rate data via controlled interference in 6G","Modulating interference powers a RedCap side channel"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000212,"raw_usage":{"total_tokens":1466,"prompt_tokens":1042,"completion_tokens":424,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":658,"completion_tokens_details":{"reasoning_tokens":350}},"tokens_in":658,"tokens_out":424,"duration_ms":3829,"temperature":1.0,"reasoning_tokens":350,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-07T14:45:49.537269+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Take a transmitter using the proposed weights, add a known mismatch between the channel used to design $\\omega_0$ and the actual secondary-user channel (for instance, a 5% norm error or a small angular drift), and measure the OOK bit-error rate as a function of $N$. If the error rate stops reaching $10^{-5}$ at the predicted $N_\\alpha$ even with the designed $\\alpha$, or if the measured sum rate drops below the primary-only rate, the central claim fails in that regime.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the Moore-Penrose pseudoinverse method used to solve the underdetermined beamforming equations for $\\omega_0$ and $\\omega_1$."},{"cited_title":"Beam index modulation wireless communication with analog beamforming,","cited_arxiv_id":null,"evidence_quote":"Baseline single-RF-chain analog-beamforming beam index modulation that the proposed scheme extends and contrasts with."},{"cited_title":"Intelligent reflecting surface assisted beam index-modulation for millimeter wave communication,","cited_arxiv_id":null,"evidence_quote":"Recent RIS-assisted beam index modulation that requires specially designed receivers, against which the no-extra-hardware claim is positioned."},{"cited_title":"Index modulation techniques for 5G wireless networks,","cited_arxiv_id":null,"evidence_quote":"Provides the index modulation concept of conveying extra information through the spatial domain that motivates embedding secondary-user data in beamformer settings."}],"review_version":1}