{"id":"83a0e490-0b50-4324-9e3e-f473abee9e79","arxiv_id":"2505.10933","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":10,"one_line_summary":"6G ISAC can be designed and evaluated across layers, from hardware and waveforms to sensing APIs, using a KPI/KVI framework demonstrated in indoor, urban, and rural settings.","lead":"This paper lays out a joint industry-academia vision for integrated sensing and communication (ISAC) in 6G, spanning radio hardware, waveforms, protocols, and privacy. It also introduces a quantitative framework that links low-level design choices to sensing accuracy, resolution, and latency, demonstrated with simulations and proof-of-concept experiments.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The paper's own text admits its quantitative core is incomplete: the promised CRLB/TPEB derivation is omitted, the promised comparison with existing sensing technologies is absent, and the results section truncates before any simulation/PoC KPIs appear.","rationale":"The reader's verdict of CONDITIONAL is reasonable, but the stress-test suggests the central claim cannot currently be verified at all. The strongest claim is quantitative: a cross-layer framework linking design parameters to KPIs and KVIs, applied via simulation and PoC to compare ISAC solutions. For this to hold, the paper must contain (a) a derivation or credible model connecting DoFs to accuracy, (b) simulation/PoC results reporting the KPIs, and (c) evidence that these results support fair comparison across ISAC solutions. None of (a)-(c) are present in the reviewed text. Section VI.B.2 explicitly says the TPEB/CRLB was derived but not included; Section III.C defers comparison with radar/camera/UWB to Section V, but Section V covers protocols/security, not comparisons; the manuscript truncates before any numerical results. The framework's qualitative links (resolution formulas, latency decomposition) are described, and the use-case tables are useful, but the claimed quantitative content is absent. Because the central evidence is missing, the correct verdict is UNVERDICTED rather than CONDITIONAL: no amount of reasoning about the present text can establish or refute the quantitative claim. The reader's OFDM-waveform concern is legitimate but secondary; it is about robustness of conclusions to waveform choice, not about whether the claimed evaluation exists. I would first verify the complete manuscript/repository before judging the framework. Credit: the paper is transparent about many gaps, links a public repository, and provides detailed use-case and hardware discussions; those parts are not at issue. The single load-bearing issue is the unavailability of the quantitative core.","tokens_in":44109,"tokens_out":5075,"duration_ms":49719,"concrete_test":"Check the linked GitHub repository (https://github.com/Hexa-X-II/ISAC-for-6G) and the complete arXiv source for: (1) a CRLB/TPEB expression for the single-point-target model of Section VI.C, and (2) KPI results (accuracy, resolution, latency) for the indoor/urban/rural simulations and the two PoC runs. Run the provided scripts for the Table XVII DoF combinations and verify that they produce a KPI table comparable to what the paper's text promises; if no CRLB/TPEB formula or no KPI output exists, the quantitative framework is not substantiated.","verdict_should_be":"UNVERDICTED","load_bearing_attack":"The central claim (Section I: 'we introduce a quantitative cross-layer framework linking design parameters to key performance and value indicators') requires the framework to actually produce quantitative KPI/KVI results. The reviewed text does not contain those results, and the omissions are acknowledged inside the paper. Section VI.B.2 states: 'Theoretical error bounds on sensing accuracy (i.e., TPEB, which is based on the CRLB) were derived but not included in the paper.' Section III.C promises a 'comparison with existing sensing technologies (e.g., radar, camera, UWB), which will be treated in Section V,' but Section V contains no such comparison. The manuscript then ends mid-sentence in Section VI.C.4, before any simulation or PoC outcome, KPI table, or KVI mapping is reported. The links shown are only definitions (range/velocity/angular resolution formulas, latency decomposition); the claimed quantitative accuracy–DoF link and the claimed application to 'compare ISAC solutions fairly' are not demonstrated in the text. This is not merely a formatting issue: it is the central evidence for the paper's main contribution. If those results and the CRLB derivation live only in the repository or in a longer version, the paper needs to carry them; if they do not exist, the central claim is unverified.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"This manuscript, developed within the Hexa-X-II project, presents a broad cross-layer vision for integrated sensing and communication (ISAC) in 6G. It reviews ISAC integration levels, use cases, KPIs, channel-model requirements, candidate radio enablers (frequency bands, RIS, NTN, massive MIMO, D-MIMO, AI), hardware architectures and impairments, higher-layer protocols and security/privacy, and then introduces a quantitative cross-layer evaluation framework based on fixed and semi-fixed degrees of freedom, resolution formulas, and a latency decomposition. The paper promises simulation and proof-of-concept results for indoor, urban, and rural scenarios at FR2/FR3, and a link from design parameters to both KPIs and KVIs.","tokens_in":44478,"tokens_out":6905,"duration_ms":66474,"significance":"If completed as promised, the proposed framework would give 6G standardization a practical tool for fair ISAC comparisons, and the qualitative synthesis is genuinely valuable: the integration-level taxonomy (Fig. 3), the enabler-by-enabler comparison tables (Tables III–VII), the treatment of PA nonlinearity and phase noise in monostatic sensing, the functional sensing-session architecture, and the STRIDE/LINDDUN threat analysis are strengths supported by extensive citations. The simulation setup is described with explicit assumptions and a public code repository link. However, the advertised quantitative contribution is not present in the reviewed text: the CRLB/TPEB accuracy bounds are stated to have been derived but omitted, and the paper ends before any KPI or KVI results are reported. The paper's value therefore currently rests on its qualitative vision and framework definition, not on the claimed quantitative evaluation.","major_comments":[{"comment":"Section VI.B.2 states that theoretical error bounds on sensing accuracy (TPEB based on the CRLB) \"were derived but not included in the paper,\" and Section VI.C.4 ends mid-sentence before any simulation or PoC KPI results, tables, or comparisons are reported. The Introduction and Contributions promise a \"quantitative cross-layer framework linking design parameters to key performance and value indicators\" and its application to simulation and PoC setups. As submitted, the quantitative core of the paper is therefore unverified: the reader is given definitions and formulas for resolution and latency, but not the accuracy-DoF link (the CRLB) nor any system-level results. This is the central claim of the paper and must be addressed, either by including the derivation and results or by explicitly reframing the contribution as a framework proposal without quantitative validation.","section":"VI.B.2, VI.C.4, I"},{"comment":"Section III.C explicitly promises a \"comparison with existing sensing technologies (e.g., radar, camera, UWB), which will be treated in Section V.\" Section V, however, covers protocols, functions, computing, and security/privacy and contains no such comparison. Since the paper motivates the framework as a way to \"compare ISAC solutions fairly,\" the missing comparison of ISAC with incumbent sensing technologies weakens the stated contribution. The comparison should be added or the promise removed.","section":"III.C, V"},{"comment":"The abstract and contribution list state that the framework links design parameters to both KPIs and KVIs. Section VI.B.2 and the evaluation setup in Section VI connect the degrees of freedom only to the sensing KPIs (accuracy, resolution, latency); no mapping from KPIs to KVIs is provided anywhere in the text. The KVI link is part of the claimed contribution and needs to be either delivered or explicitly deferred.","section":"I, VI.B.2"}],"minor_comments":[{"comment":"In the latency decomposition, the symbol T_proc^prop is used both for the propagation delay between sensing and processing nodes and for the computation time C/F; the latter should be T_proc^comp. The current notation makes the equation internally inconsistent.","section":"VI.B.2"},{"comment":"Table XVIII lists \"300 GLFOP\" for the core computational power; this appears to be a typo for \"GFLOP\", consistent with the other entries and with the units stated in the text.","section":"VI.B.1"},{"comment":"The word \"applicaitons\" should be \"applications\".","section":"V.C.2"},{"comment":"The term \"device class RHDRBL\" is introduced without definition or explanation; it should be defined or removed.","section":"V.C.4"},{"comment":"In the many-to-one multistatic sensing enumeration, the two options are both labelled \"(ii)\"; the second should be \"(i)\" (or the list should be renumbered).","section":"III.C.4"},{"comment":"In the Level 2 bullet, \"analog-to-digital converters (ADCs), analog-to-digital are utilized\" should read \"analog-to-digital converters are utilized\".","section":"II.A"}],"recommendation":"major_revision","confidential_remarks":"The manuscript appears to be an incomplete version of a larger project document: the text ends mid-sentence at Section VI.C.4, and the paper itself acknowledges that the CRLB derivation and some results are not included. Before further review, the authors should be asked to provide a complete manuscript containing the missing evaluation section or to revise the stated contributions accordingly. If the simulation/PoC results and the CRLB derivation exist in the public repository or in Hexa-X-II deliverables, integrating them into the paper would likely resolve the main concern. The broad author list and project context are not themselves a problem, but the paper's fit as a journal article depends on the quantitative section being complete."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Quick take: this is a genuinely useful cross-layer synthesis of ISAC for 6G, but the paper's central quantitative claim is not currently verified. The abstract promises a framework applied to simulation and proof-of-concept setups, yet the text stops short of showing any KPI results, and the paper itself admits the TPEB/CRLB derivation was left out.\n\nWhat it does well: the qualitative content is strong. The tables mapping frequency bands, RIS, NTN, massive MIMO, and D-MIMO to coverage, resolution, mobility, propagation, and hardware aspects are practical and will be handy in standardization discussions. The treatment of PA nonlinear distortion and phase noise in monostatic sensing, including the range-correlation effect and delayed-LO mitigation, is specific and well-grounded. The higher-layer architecture (SeMF, SCF, SPF, SUSF, SPCTM, Sensing Store) and the STRIDE/LINDDUN security-privacy analysis are more detailed than most ISAC papers. The authors also state their gaps explicitly, which I appreciate.\n\nThe soft spot is exactly where the abstract points. Section VI.B.2 says the accuracy bounds were derived but not included. Section III.C promises a comparison with existing sensing technologies \"in Section V,\" but Section V does not contain it. The manuscript ends mid-sentence in VI.C.4, before any simulation or PoC KPI appears. So the quantitative cross-layer framework is currently a framework plus resolution and latency formulas, not a demonstrated evaluation tool. That is a load-bearing gap, but a fixable one: the GitHub repo is referenced, so the authors could add a few KPI tables, a short description of the CRLB setup, and at least one accuracy-versus-resource plot. The waveform assumption (OFDM/DFTS-OFDM) is a stated bet rather than a hidden one; fine, but it makes the quantitative insights conditional on that choice.\n\nProportionate verdict: the qualitative content stands on its own, and the missing numbers do not invalidate the synthesis. As submitted, though, the paper overpromises. I would accept it for peer review, because the framework and the enabler tables deserve refereeing, and the missing pieces are checkable via the repository. I would also push for major revision to make the quantitative contribution visible. If the repository lacks the results, the authors should recast the contribution as a qualitative framework plus a research agenda.\n\nWho benefits: standardization engineers and ISAC researchers wanting a cross-layer map. I would bring it to a reading group, and I would cite the enabler tables in my own work. Recommendation: send it to serious referees, but condition acceptance on the authors showing the actual numbers.","headline":"A broad, well-sourced ISAC vision paper whose promised quantitative evaluation is missing from the text; worth a serious referee, but only after the authors actually show the numbers.","tokens_in":45178,"tokens_out":2369,"would_cite":true,"duration_ms":25984,"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":"This paper introduces a quantitative cross-layer framework that links 6G ISAC design choices to sensing accuracy, resolution, latency, and high-level value indicators.","keywords":["integrated sensing and communication","6G","cross-layer evaluation","sensing KPI","key value indicators","OFDM","sensing latency","proof of concept"],"falsifier":"Measure range error and range-Doppler noise floor on a real OFDM monostatic ISAC prototype at 30 GHz or 300 GHz with a corner reflector at a known distance, comparing against the paper's phase-noise and power-amplifier distortion curves across transmitter back-off levels; the framework's hardware link would be falsified if measured errors systematically exceed the predicted values by more than the model variance, or if the predicted preference for OFDM over single-carrier reverses.","tokens_in":43940,"feed_emoji":"📡","tokens_out":7748,"duration_ms":72592,"temperature":0.7,"pith_summary":"This paper argues that integrated sensing and communication (ISAC) in 6G can be designed and compared fairly only through a cross-layer framework that connects low-level radio choices to sensing accuracy, resolution, and latency, and then to high-level value indicators. It proposes such a framework, built on fifteen degrees of freedom spanning radio configuration (carrier frequency, antennas, subcarriers, OFDM symbols, power) and network processing (where and with what power sensing data is processed), and applies it in simulation and proof-of-concept experiments for indoor, urban, and rural use cases. The authors' central claim is that physical-layer trade-offs, such as waveform choice versus power-amplifier distortion or bandwidth versus coverage, propagate upward and can be quantified so that resource allocation between sensing and communication becomes a system-level decision rather than a per-layer one. A sympathetic reader would care because, if the framework holds, 6G standardization gains a concrete way to evaluate ISAC solutions before deployment, including how sensing latency and accuracy respond to processing location and computational power.","feed_headline":"A cross-layer framework makes 6G sensing trade-offs computable","feed_subtitle":"Sensing accuracy, resolution, and latency become predictable from radio and compute choices, enabling fair 6G ISAC design.","key_machinery":"The central machinery is the cross-layer evaluation framework itself: a parameterized mapping from fifteen degrees of freedom to sensing KPIs. Its operational core is a set of quantitative relations—the resolution formulas $\\Delta R=c/(2B)$, $\\Delta v=\\lambda/(2T_{\\mathrm{Tx}})$, and $\\Delta\\phi=0.89\\lambda/D$; the total latency model $T_{\\mathrm{total}}=T_{\\mathrm{Tx}}+T_{\\mathrm{prop}}+T_{\\mathrm{proc}}$ with $T_{\\mathrm{proc}}=T_{\\mathrm{proc}}^{\\mathrm{Tx}}+T_{\\mathrm{proc}}^{\\mathrm{prop}}+T_{\\mathrm{proc}}^{\\mathrm{comp}}$, where the data-transfer and compute terms depend on processing-location proximity, communication bandwidth, and computational power; and a Cramér–Rao-based accuracy link that connects all twelve radio-layer degrees of freedom. This machinery carries the argument by turning the qualitative claim that physical-layer choices propagate upward into numbers that can be compared across use cases, frequency bands, and processing placements.","core_discovery":"The paper's central claim is that ISAC should be evaluated as a system, not layer by layer, and that a quantitative cross-layer framework can make this evaluation concrete. The framework links fifteen degrees of freedom—twelve radio-layer choices (carrier frequency, number of antennas at base station and user equipment, subcarrier spacing, number of OFDM symbols and subcarriers, transmit power, precoder/combiner design, base station locations, and number of RF chains) and three networking-layer choices (processing location, communication bandwidth to the processor, and computational power)—to sensing KPIs of accuracy, resolution, and latency. Resolution is predicted from closed-form formulas ($\\Delta R=c/(2B)$, $\\Delta v=\\lambda/(2T_{\\mathrm{Tx}})$, $\\Delta\\phi=0.89\\lambda/D$); latency is decomposed into transmission, propagation, and processing terms, where processing includes moving the sensing data to a compute node and performing the estimation there; accuracy is tied to the radio degrees of freedom through a Cramér–Rao bound. The paper then exercises the framework on three scenarios: indoor bistatic human detection, urban bistatic intersection hazard detection, and rural monostatic highway hazard detection, using both simulation and proof-of-concept hardware, and reports how the degrees of freedom move the KPIs.","pith_inferences":["Beyond the paper: since the latency model separates transmission, propagation, and compute time, it could be turned into a placement rule—for a target speed of 10 m/s and a required accuracy of 1 m, the total sensing latency budget is 100 ms, which immediately constrains how far from the radio the sensing processor can live.","Beyond the paper: the KVI side of the framework is left qualitative; a natural next step would be to attach utility or revenue to accuracy-latency packages for each use case, making the KPI-to-KVI link numerically testable.","Beyond the paper: the same framework structure could be applied to non-OFDM waveforms or to sub-THz carriers once those are available, since the resolution and latency formulas are waveform-general even though the distortion results are not."],"forward_implications":["A fair comparison of ISAC solutions becomes possible: two designs (e.g., FR3 versus FR2, bistatic versus monostatic) can be evaluated on the same KPI axes of accuracy, resolution, and latency across indoor, urban, and rural scenarios.","Sensing latency becomes partly a networking decision: the model predicts that processing location and computational power, not just radio parameters, set total latency, and that for a target moving at 10 m/s a 100 ms latency contributes 1 m of error independent of estimator accuracy.","Band selection gets quantitative guidance: the range, velocity, and angular resolution formulas show that FR2 and FR3 choices trade coverage and mobility against resolution, with sub-THz offering centimeter-scale range resolution but severe coverage limits.","The waveform discussion concludes that OFDM remains preferable to single-carrier for sensing under power-amplifier nonlinearity in most operating regions, because single-carrier's lower distortion is outweighed by noise enhancement during frequency-domain division."],"supporting_citations":[{"why":"Defines the key value indicators used as the high-level endpoint of the cross-layer framework.","marker":"[16]"},{"why":"Earlier work that this paper builds on and extends with more detail and depth.","marker":"[17]"},{"why":"Earlier work that this paper builds on and extends with more detail and depth.","marker":"[18]"},{"why":"Supplies the end-to-end 6G system architecture onto which the ISAC enablers are mapped.","marker":"[75]"},{"why":"Introduces the sensing management function and the ISAC enablers including the OFDM frame-structure assumption.","marker":"[76]"},{"why":"Provides the power-amplifier distortion comparison between OFDM and single-carrier that drives the waveform discussion.","marker":"[126]"},{"why":"Gives the delay-dependent phase-noise model used in the monostatic ranging analysis.","marker":"[170]"},{"why":"Introduces the target-channel versus background-channel separation used for ISAC channel modeling and simulation.","marker":"[57]"},{"why":"Supplies the accuracy, velocity, and latency requirements assumed for the urban intersection use case.","marker":"[218]"}],"fun_headline_variants":["Cross-layer framework turns 6G sensing trade-offs into numbers","6G ISAC design gets a quantitative cross-layer tool","From radio choices to sensing KPIs: a 6G ISAC framework","Predicting 6G sensing resolution, latency, and accuracy","Cross-layer framework makes ISAC trade-offs computable"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The framework's quantitative results rest on the premise that 6G sensing waveforms will be OFDM or its precoded variant DFTS-OFDM; if a different waveform family is standardized, the resolution, power-amplifier distortion, and resource-allocation conclusions would need to be re-derived.","fun_headline_variants_meta":{"raw":{"variants":["Cross-layer framework turns 6G sensing trade-offs into numbers","6G ISAC design gets a quantitative cross-layer tool","From radio choices to sensing KPIs: a 6G ISAC framework","Predicting 6G sensing resolution, latency, and accuracy","Cross-layer framework makes ISAC trade-offs computable"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000201,"raw_usage":{"total_tokens":1408,"prompt_tokens":1002,"completion_tokens":406,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":618,"completion_tokens_details":{"reasoning_tokens":320}},"tokens_in":618,"tokens_out":406,"duration_ms":3950,"temperature":1.0,"reasoning_tokens":320,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-15T21:00:40.735736+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Measure range error and range-Doppler noise floor on a real OFDM monostatic ISAC prototype at 30 GHz or 300 GHz with a corner reflector at a known distance, comparing against the paper's phase-noise and power-amplifier distortion curves across transmitter back-off levels; the framework's hardware link would be falsified if measured errors systematically exceed the predicted values by more than the model variance, or if the predicted preference for OFDM over single-carrier reverses.","supporting_citations":[{"cited_title":"Service requirements for integrated sensing and communication (3GPP TS 22.137 version 19.0.0 Release 19),","cited_arxiv_id":null,"evidence_quote":"Supplies the accuracy, velocity, and latency requirements assumed for the urban intersection use case."}],"review_version":1}