{"id":"913354cb-e9ec-4fd1-b553-7d2b73bc0013","arxiv_id":"1908.11016","paper_version":2,"verdict":"CONDITIONAL","confidence":"HIGH","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":4,"one_line_summary":"A joint waveform and receive-filter optimization, solved by sequential convex programming, lets a hybrid radar maintain high SINR despite timing uncertainty in the co-channel communication signal.","lead":"Radar engineers can now jointly shape the transmit waveform and receive filters of a hybrid active-passive radar so it stays effective even when the passive signal arrives with uncertain timing. This is a method for letting radar sensing and communication signals share the same frequency band with less mutual interference.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Timing-uncertainty robustness is demonstrated only for integer sample shifts; the continuous/fractional delay case is outside the Jk model and the MM/WS guarantee.","rationale":"The reader's conditional verdict identifies the integer-shift timing model as the weakest assumption, and my stress-test converges to the same point. The paper is a competent contribution: the signal model is clearly laid out, the MM and WS formulations are reasonable, the SDR/SCP solution procedures are standard for this class of nonconvex problems, and the numerical comparisons show meaningful gains under the assumed model. However, the paper's headline claim of robustness to timing uncertainty is exactly as strong as the timing model, and that model only covers integer sample shifts. Real propagation delays and synchronization errors are not quantized to the sampling grid, so a fractional delay δ ∈ (0,1) falls outside the Jk representation. Since the designs optimize only over the finite set NK and Fig. 7 tests only integer k, the reported MM flatness and WS robustness do not bound the continuous worst case. This is a load-bearing issue for the central claim, but it is a limitation that can be addressed by extending the model or by explicitly narrowing the claim to integer-sample timing offsets. The paper also lacks code, data, and error bars, and the SDR/SCP methods have no global convergence guarantees, yet these are secondary: the most important correctness risk is the timing-quantization assumption. Because the reader already conditioned acceptance on such concerns, the verdict does not need to change; the condition should be made explicit as a required revision.","tokens_in":19082,"tokens_out":4392,"duration_ms":49989,"concrete_test":"Run a fractional-delay robustness experiment using the Section V parameters: design sr, wr, wc via the MM and WS algorithms with K=3, then evaluate the actual SINR when the true relative delay is (k+δ)Ts for k ∈ {-3,...,3} and δ sampled uniformly in [0,1), constructing the IO signal directly with Hδ whose entries are g(pTs - iTc - δTs) instead of using JkHb. Compare the worst-case SINR over the continuous interval with (i) the integer-only worst case reported in Fig. 7 and (ii) the active-only and passive-only baselines. If the continuous worst-case SINR drops by more than 2-3 dB below the integer worst case, or falls to the baseline level, the robustness claim in the abstract and Section V is overstated.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central robustness claim rests on the integer-shift model in Eqs. (19)-(21), where the timing offset is taken to be an integer k and the shifted IO signal is written as sc = JkHb. In a real system the relative delay between the communication and radar waveforms, caused by target/IO location error or synchronization error, is generally not an integer multiple of the sampling interval Ts. For a delay (k+δ)Ts with δ∈(0,1), the correctly sampled IO vector is not JkHb: the pulse-shaping matrix H must be rebuilt with entries g(pTs - iTc - δTs), so H itself becomes delay-dependent. The MM problem (23) and WS problem (24) optimize over the finite integer set NK, and Fig. 7 evaluates the designs only at integer k ∈ {-K,...,K}. Consequently, the MM design's 'uniform SINR over the timing uncertainty interval' and the WS design's reported robustness are statements about the discrete model, not about continuous timing uncertainty. If the true fractional delay is not one of the tested integer shifts, no performance guarantee is provided, and the worst case over the continuous uncertainty interval can be worse than the integer-only worst case. A secondary fragility is the assertion Rb = I 'without loss of generality' after Eq. (22): for correlated IO symbols or non-unit symbol powers this is not absorbed by rescaling γc, and the interference covariance changes in a way the designs do not address. The paper's sampling assumption (Ts very small) does not remove fractional delay unless sub-sample delay alignment is explicitly modeled, and no such mechanism appears in the signal model.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper considers a co-channel hybrid active-passive radar in which a monostatic active radar and a non-cooperative illuminator of opportunity (IO) operate in the same frequency band. The received signal contains target echoes from both the active and passive paths, and the radar jointly designs its transmit waveform and two receive filters to maximize the output SINR, assuming the IO modulation format and pulse shape are known but the data symbols are not. Timing uncertainty is modeled as an integer sample shift k in a shift matrix Jk (Eqs. (19)-(21)), and two design criteria are proposed: a max-min (MM) design that maximizes the worst-case SINR over k in {-K,...,K}, and a weighted-sum (WS) design that maximizes a weighted average of the SINR over this set. Both problems are nonconvex and are solved by alternating sequential convex programming (SCP) and semidefinite relaxation (SDR) with randomization, with a simplified closed-form receiver update for the K=0 case. Numerical simulations show convergence of the algorithms, SINR improvements over active-only and passive-only baselines, robustness of the MM and WS designs over integer delays, and better detection probability for the hybrid designs.","tokens_in":19393,"tokens_out":4177,"duration_ms":43844,"significance":"If the results hold, the paper makes a useful contribution to co-channel hybrid radar design by explicitly accounting for timing uncertainty between the active and passive illuminators. The signal model is clearly formulated, the optimization framework is technically sound as a heuristic, and the simulation study is reasonably extensive, including convergence checks, SINR contours, timing-robustness plots, and detection probability. Strengths include the explicit derivation of the SINR expressions, the algorithmic detail, and the falsifiable numerical predictions. The main limitation is that the robustness claims are established only for integer sample shifts, which narrows the scope of the headline result. The lack of convergence guarantees for the alternating SCP/SDR procedure is a secondary gap that should be acknowledged.","major_comments":[{"comment":"The timing uncertainty model is restricted to integer sample shifts. For a fractional delay (k+δ)Ts with δ in (0,1), the sampled IO signal is not JkHb, because the pulse-shaping matrix H in Eq. (14) would need to be rebuilt with samples g(pTs - iTc - δTs), making H itself delay-dependent. Consequently, the MM design's claimed 'uniform output SINR over the timing uncertainty interval' and the WS design's reported robustness are statements about the discrete set {-K,...,K} only. Fig. 7 also evaluates the designs only at integer k. The paper should either clearly scope all robustness claims to integer sample shifts (and qualify the abstract and introduction accordingly), or extend the signal model to continuous delays and re-evaluate the worst-case performance over the continuous uncertainty interval.","section":"Section III-B, Eqs. (19)-(21); Fig. 7"},{"comment":"The statement 'Without loss of generality, we assume Rb = I' is only valid when the communication symbols are uncorrelated and have equal power, i.e., Rb = σ²I, because a scalar factor can be absorbed into γc. For correlated symbols or non-uniform symbol powers, the interference covariance is not a scaled identity, and the subsequent MM/WS designs do not address the resulting covariance structure. Please replace the WLOG claim with an explicit assumption of uncorrelated equal-power symbols, or justify the reduction for arbitrary Rb.","section":"Section III-C, after Eq. (22)"},{"comment":"The alternating SCP/SDR procedure is heuristic: no convergence guarantee is provided for the outer loop or the SCP inner loop, and the randomization step produces approximate rank-one solutions. The stopping criterion is based on observed SINR improvement, and Fig. 5 demonstrates empirical convergence for one configuration. This is acceptable as a numerical method, but the paper should state clearly that only local/empirical convergence is claimed, and ideally provide a short discussion of conditions under which the SCP updates are guaranteed to improve the objective (e.g., monotonicity properties of the inner problems). Without this, the reported SINR values should be interpreted as achievable by the specific algorithm, not as globally optimal SINR values.","section":"Section IV, Algorithms 2 and 3"}],"minor_comments":[{"comment":"In the last term of Eq. (34), the expectation should read E{ξ^H (γc Jk H H^H Jk^H + I) ξ}, with a conjugate transpose on the first ξ; the current expression omits the Hermitian transpose and is dimensionally inconsistent.","section":"Eq. (34)"},{"comment":"The x-axis label 'real relative delay k' suggests a continuous quantity, but only integer sample shifts are evaluated. Please state in the caption that k is an integer sample index, or plot markers at the evaluated integer points.","section":"Fig. 7"},{"comment":"A brief explanatory sentence after Step 2 of Algorithm 1 would help: the update of λ uses the ratio f_k/g_k with f_k including the κ_k term, so that κ_k is not double-counted in the subsequent convex problem (30).","section":"Algorithm 1"},{"comment":"The single-pulse assumption and the neglect of Doppler are stated, but the detection results in Section V-C are also for a single pulse; a sentence connecting the single-pulse SINR design to the energy-detector detection simulation would improve clarity.","section":"Section II, Remark 1"}],"recommendation":"major_revision","confidential_remarks":"The paper is within the scope of a signal processing journal and the topic is timely. The central algorithmic framework is reasonable, but the integer-shift timing model is a load-bearing simplification for the robustness claims, and the 'without loss of generality' statement about Rb should be corrected. The lack of convergence guarantees is a standard limitation in this line of work and can be addressed by a careful statement of scope. I see no citation-pattern issues or novelty concerns; the related work is cited appropriately."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"This is a competent, clearly written paper that does something genuinely new in a modest way: it combines co-channel hybrid active-passive radar with timing-uncertainty-robust joint waveform and receiver design. The Jk shift-matrix model and the max-min (MM) and weighted-sum (WS) criteria are not in the cited literature, and the resulting optimization problems, while assembled from known tools like SDR, Dinkelbach, quadratic transform, and SCP, are applied to a problem that matters for spectrum sharing. The signal model and SINR derivations are internally consistent, and the simulations support the claimed gains. The robustness test in Fig. 7, which evaluates designs at delays not used in the optimization, provides real external support rather than pure curve-fitting. Citation practice looks honest, including self-citations to prior hybrid radar work.\n\nThe soft spots are real but not fatal. The biggest one is the integer-shift assumption. The model treats timing uncertainty as an integer sample shift k, with sc = JkHb. In practice the relative delay is continuous, and for a fractional delay the pulse-shaping matrix H itself becomes delay-dependent; the correctly sampled IO vector is not JkHb for any integer k. The MM and WS designs therefore optimize over a finite discrete set, and Fig. 7 only tests integer k. So the claim of robustness over a timing uncertainty interval is stronger than what the model supports. The paper should either handle fractional delay explicitly (e.g., a finer grid or a parameterized H(delay)) or carefully rephrase the robustness claim to be about integer sample offsets. This is a scope issue, not a sign that the central method is wrong within its model.\n\nTwo smaller issues. First, the \"without loss of generality, Rb = I\" after Eq. (22) is only fine for i.i.d. unit-power symbols; correlated symbols or non-unit powers change the interference covariance in a way the designs do not address. Second, the optimization machinery is heuristic: no global optimality or convergence proof for the alternating SCP/SDR loops, and randomization gives approximate rank-one solutions. The numerical convergence plots help, but there is no code or data, and the figures have no error bars. These are common in this literature and should be flagged, not disqualifying.\n\nWho is this for? Researchers working on hybrid radar, passive sensing, and radar-communication coexistence. They will find a useful application-oriented method and a clear baseline for comparison. It is not a breakthrough framework, but it is a solid new application with correct derivations and honest simulations.\n\nI would engage with this paper in review. The integer-shift limitation is fixable with revised claims or a fractional-delay extension, and the rest of the work holds up. A serious editor should send it to peer review, with the expectation of a revision that either narrows the robustness claim or extends the model.","headline":"Solid within-subfield contribution: new robust joint waveform/filter designs for co-channel hybrid active-passive radar under timing uncertainty, with a real but fixable gap around fractional timing shifts.","tokens_in":817,"tokens_out":916,"would_cite":true,"duration_ms":26554,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"This paper claims that a co-channel hybrid active-passive radar, with jointly designed radar waveform and receive filters, achieves substantially higher output SINR and better detection than either active-only or passive-only radar, while…","keywords":["hybrid active-passive radar","joint waveform and receiver design","timing uncertainty","co-channel spectrum sharing","output SINR maximization","max-min optimization","weighted-sum optimization","semidefinite relaxation"],"falsifier":"Measure the output SINR and detection probability of the $K=3$ designs when the true relative delay is half a sample, or any non-integer shift. A substantial drop below the designed worst-case SINR, or a larger miss-probability loss than seen for integer shifts, would show the integer-shift model is the limiting assumption. A separate test with correlated IO symbols, i.e. $\\mathbf{R}_b \\neq \\mathbf{I}$, would show how sensitive the design is to the uncorrelated-symbol assumption.","tokens_in":18888,"feed_emoji":"📡","tokens_out":6028,"duration_ms":56418,"temperature":0.7,"pith_summary":"The paper considers a radar that both transmits its own probing waveform and passively listens to a communication transmitter serving as an illuminator of opportunity in the same frequency band. It claims that jointly designing the radar waveform and two receive filters can considerably raise the output SINR compared with using either active-only or passive-only sensing, and that the design can be made robust to unknown relative timing between the radar and communication signals. To handle timing uncertainty, it proposes a max-min criterion that guards the worst delay and a weighted-sum criterion that averages over possible delays, both solved as nonconvex optimization problems. When timing uncertainty vanishes, the two designs become identical and the receive-filter updates reduce to closed forms. The intended payoff is spectral and energy efficiency without giving up detection performance.","feed_headline":"Hybrid radar design lifts SINR over active-only or passive-only","feed_subtitle":"A max-min and a weighted-sum design keep detection robust when the communication signal's timing is uncertain.","key_machinery":"The load-bearing construction is the discrete-time model of the illuminator-of-opportunity signal as a known pulse-shaping matrix $H$ multiplying unknown uncorrelated symbols $\\mathbf{b}$, followed by an integer sample-shift matrix $J_k$: $\\mathbf{s}_c = J_k H \\mathbf{b}$. The shift index $k$ carries all location- and synchronization-induced timing uncertainty, and the output SINR for each $k$ splits into an active-branch and a passive-branch generalized Rayleigh quotient. On top of this model, the solution machinery alternates among the two receive filters and the radar waveform, using semidefinite relaxation with randomization, a fractional-programming iteration for worst-case ratio problems, a quadratic-transform iteration for sums of ratios, and sequential convex programming with a first-order approximation for the waveform update; when $K=0$ both design criteria reduce to a common problem whose receive-filter updates are closed-form.","core_discovery":"The central claim is that a co-channel hybrid active-passive radar can treat the communication signal as a second useful illumination rather than only as interference, provided the radar waveform and receive filters are optimized jointly. The paper models the communication waveform as a known pulse shape multiplied by unknown, uncorrelated symbols, with the timing mismatch represented as an integer sample shift $k$. It derives the output SINR as a sum of an active-path and a passive-path SINR, then formulates the design as either maximizing the worst-case SINR over $k$ in a bounded interval or maximizing a weighted sum of SINR values over that interval, subject to a transmit power constraint. Because the resulting problems are nonconvex, the paper proposes alternating optimization using semidefinite relaxation with randomization, fractional programming for worst-case ratios, a quadratic-transform method for sums of ratios, and sequential convex programming for the waveform update. Numerical results show the hybrid designs outperform active-only and passive-only systems, with the max-min design giving a uniform SINR floor across the assumed uncertainty interval and the weighted-sum design losing little at the nominal delay.","pith_inferences":["The integer-shift timing model could be extended to fractional delays by replacing $J_k$ with a fractional-delay filter; the same alternating optimization framework would likely carry over, but the paper does not test this.","The weighted-sum criterion's weights could encode a prior distribution over delay, turning the design into expected-SINR maximization; the paper only simulates uniform weights.","The diversity-order gain visible in the detection curves suggests an analytic diversity analysis of hybrid active-passive detection is a natural follow-up not pursued here.","Because the paper assumes radar-to-communication interference is negligible under directive transmission, the design is tailored to that spectrum-sharing regime; relaxing this assumption would couple the radar waveform design to communication-receiver constraints."],"forward_implications":["A hybrid radar can trade active transmit power against illuminator-of-opportunity strength while holding target SINR fixed, so co-channel operation offers an energy-efficiency lever.","A max-min design guarantees an SINR floor across the entire assumed delay uncertainty interval, whereas a design that ignores timing uncertainty loses SINR when the real delay is nonzero.","The weighted-sum design with a nonzero uncertainty bound loses little at the nominal delay, so robustness can be obtained without sacrificing the ideal no-uncertainty performance.","Jointly optimizing the radar waveform and receive filters outperforms optimizing only the receive filters with a fixed radar waveform in the simulated scenarios.","Detection simulations show the hybrid scheme has a steeper miss-probability slope than active-only or passive-only sensing, indicating a spatial diversity gain from using both illumination paths."],"supporting_citations":[{"why":"Introduces the known-communication-signal-form model used to build $\\mathbf{s}_c = H\\mathbf{b}$ and motivates exploiting partial IO knowledge.","marker":"[8]"},{"why":"Defines the hybrid active-passive radar concept that this paper extends to co-channel operation with timing uncertainty.","marker":"[11]"},{"why":"Provides the co-channel MIMO gain result showing the benefit of using communication transmitters as extra illuminators.","marker":"[19]"},{"why":"Supplies the semidefinite relaxation and randomization approach used to convert each nonconvex rank-one filter update into a convex problem.","marker":"[40]"},{"why":"Supplies the fractional-programming iteration used to solve the worst-case SINR subproblem in the max-min design.","marker":"[41]"},{"why":"Provides the sequential convex programming framework used for the radar waveform update.","marker":"[44]"},{"why":"Provides the quadratic-transform method used to solve the sum-of-ratios subproblem in the weighted-sum design.","marker":"[47]"},{"why":"Supplies the energy-detector combining rule and the diversity-gain interpretation used in the detection simulations.","marker":"[48]"}],"fun_headline_variants":["Hybrid radar outperforms active or passive alone","Joint waveform-receiver design handles timing uncertainty","Max-min and weighted-sum rules for robust radar SINR","Co-channel hybrid radar lifts SINR under timing jitter","Radar design beats active-only despite unknown delay"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The design assumes the timing mismatch is an integer number of sampling intervals and that the IO pulse shape and modulation are known while the data symbols are uncorrelated; if the actual delay is fractional or the communication signal's statistics differ, the shift-matrix model and the optimized designs are no longer matched to the received signal.","fun_headline_variants_meta":{"raw":{"variants":["Hybrid radar outperforms active or passive alone","Joint waveform-receiver design handles timing uncertainty","Max-min and weighted-sum rules for robust radar SINR","Co-channel hybrid radar lifts SINR under timing jitter","Radar design beats active-only despite unknown delay"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.00024,"raw_usage":{"total_tokens":1535,"prompt_tokens":980,"completion_tokens":555,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":596,"completion_tokens_details":{"reasoning_tokens":481}},"tokens_in":596,"tokens_out":555,"duration_ms":6192,"temperature":1.0,"reasoning_tokens":481,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T10:26:51.666935+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Measure the output SINR and detection probability of the $K=3$ designs when the true relative delay is half a sample, or any non-integer shift. A substantial drop below the designed worst-case SINR, or a larger miss-probability loss than seen for integer shifts, would show the integer-shift model is the limiting assumption. A separate test with correlated IO symbols, i.e. $\\mathbf{R}_b \\neq \\mathbf{I}$, would show how sensitive the design is to the uncorrelated-symbol assumption.","supporting_citations":[{"cited_title":"Improved detection performance for passive radars exploiting known communication signal form,","cited_arxiv_id":null,"evidence_quote":"Introduces the known-communication-signal-form model used to build $\\mathbf{s}_c = H\\mathbf{b}$ and motivates exploiting partial IO knowledge."},{"cited_title":"Joint transmit and receive beamforming for hybrid activepassive radar,","cited_arxiv_id":null,"evidence_quote":"Defines the hybrid active-passive radar concept that this paper extends to co-channel operation with timing uncertainty."},{"cited_title":"Performance gains from cooperative MIMO radar and MIMO communication systems,","cited_arxiv_id":null,"evidence_quote":"Provides the co-channel MIMO gain result showing the benefit of using communication transmitters as extra illuminators."},{"cited_title":"Semidefinite relaxation of quadratic optimization problems,","cited_arxiv_id":null,"evidence_quote":"Supplies the semidefinite relaxation and randomization approach used to convert each nonconvex rank-one filter update into a convex problem."},{"cited_title":"Local convergence of sequential convex programming for nonconvex optimization,","cited_arxiv_id":null,"evidence_quote":"Provides the sequential convex programming framework used for the radar waveform update."},{"cited_title":"Fractional programming for communication systems part I: Power control and beamforming,","cited_arxiv_id":null,"evidence_quote":"Provides the quadratic-transform method used to solve the sum-of-ratios subproblem in the weighted-sum design."},{"cited_title":"MIMO radar with widely separated antennas,","cited_arxiv_id":null,"evidence_quote":"Supplies the energy-detector combining rule and the diversity-gain interpretation used in the detection simulations."}],"review_version":1}