{"id":"7ff73b5a-8c04-46c9-8391-7eb161f65ec8","arxiv_id":"2501.01200","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":4.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"A survey and tutorial that classifies redundancy mitigation strategies for vehicular collective perception into object inclusion filtering, data format optimization, and frequency management, and contrasts the 2019 and 2023 ETSI specifications.","lead":"This paper reviews and categorizes the many ways researchers have proposed to reduce redundant messages in vehicle-to-vehicle cooperative perception, where cars broadcast what their sensors detect. It compares the 2019 and 2023 European standards and groups all approaches into three buckets: filtering which objects to report, compressing the data format, and reducing how often messages are sent.","discovery_kind":"review","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The three-category taxonomy is non-disjoint and inconsistently applied, so the central claim that it provides a clear framework is not fully supported.","rationale":"The reader's concern about the undocumented negative claim that none of the new VoI-based measures have been evaluated is legitimate and worth resolving, but it is not the most load-bearing issue for the paper's central claim. The survey's stated contribution is the taxonomy as a clear framework for comparing and extending the literature. That contribution is undercut by internal inconsistencies: the taxonomy is neither disjoint nor exhaustive as applied. Frequency and Content Management changes both frequency and content; MCO is explicitly outside the taxonomy; and Where2comm, which shares compressed feature maps rather than filtering detected objects, is placed under Object Inclusion Filtering. These are not merely cosmetic labeling issues because the taxonomy is the paper's main organizing contribution and is used to structure Section IV and Fig. 9. The paper does have real strengths: it provides a detailed comparison of the 2019 and 2023 ETSI specifications, a useful compendium of performance results, and a thoughtful discussion of open problems. Those strengths justify keeping the survey as a conditional reference, but the taxonomy should be revised or explicitly framed as a heuristic with overlapping categories before the central claim is accepted as stated.","tokens_in":34833,"tokens_out":5310,"duration_ms":54593,"concrete_test":"Extract every entry in Fig. 9 and apply the Section IV definitions mechanically: (1) does the method decide which objects or regions are included in a message, (2) does it compress or alter the data format per included object, and (3) does it change message generation frequency? Record all categories that apply for each entry. If any entry is multi-labelled or if Where2comm, ETSI Frequency and Content Management, and MCO cannot be assigned uniquely, the claim of three distinct and exhaustive classes needs qualification.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The survey's central value proposition is that the proposed taxonomy — object inclusion filtering, data format optimisation, frequency management — gives a clear and complete framework for understanding the state of the art (Section I, fifth paragraph). That claim requires each approach to be assignable to exactly one well-defined category, but the paper's own assignments contradict this. Section II-E defines the ETSI 'Frequency and Content Management' measure as jointly selecting both the CPM generation interval and the number of perceived objects/regions to include, yet Fig. 9 places it only under Frequency Management even though it is also a content-inclusion decision. The authors also concede in Section IV that Multi-Channel Operation 'does not strictly fit the classification,' and Section IV-D introduces 'Infrastructure-Based Redundancy Mitigation Mechanisms' as an additional category outside the three-part taxonomy. Most concretely, Where2comm is grouped under Object Inclusion Filtering in Fig. 9, while Section IV-A6 describes it as transmitting 'a spatially sparse feature map' to reduce raw sensor data, which matches the Data Format Optimisation definition instead. If assignments depend on editorial judgment rather than the stated definitions, the taxonomy cannot serve as the stable, common structure that the central claim promises.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"This manuscript surveys and compares redundancy mitigation mechanisms for the ETSI Collective Perception Service, contrasting the 2019 ETSI TR 103 562 rules with the 2023 ETSI TS 103 324 Value-of-Information-based measures. It reviews literature quantifying the communication burden of CPMs, evaluates studies of the ETSI redundancy rules, and proposes a three-category taxonomy of academic mitigation approaches: object inclusion filtering, data format optimisation, and frequency management. The paper also discusses simulation environments for CPS evaluation and closes with open research issues such as adaptive threshold tuning, multi-channel operation, semantic communication, and the interaction between redundancy mitigation and security, mapping, and trust.","tokens_in":35012,"tokens_out":6237,"duration_ms":66445,"significance":"If the taxonomy and the survey's organization are made internally consistent, this will be a useful reference for researchers entering the CPS redundancy-mitigation area. The paper's strengths include a detailed side-by-side account of ETSI TR 103 562 vs TS 103 324, a clear summary of the eight VoI methods in Table II, a broad and current citation base, and an unusually thoughtful open-issues section covering cybersecurity, LDM/BEV, belief/certainty, and semantic communication. No derivations or code are expected for a survey, and the paper does not overclaim original technical results. However, the central taxonomic claim needs tightening, and the strong assertion that none of the new VoI measures have been evaluated needs either a documented search procedure or a more carefully qualified formulation.","major_comments":[{"comment":"The paper's central claim that the proposed three-part taxonomy \"provides a clear framework for understanding and extending the state of the art\" is not fully supported, because the categories are not applied consistently. Section II-E describes ETSI's Frequency and Content Management as jointly selecting both the CPM generation interval and the number of perceived objects/regions to include, yet Fig. 9 places it only under Frequency Management, ignoring its content-inclusion component. Where2comm is grouped under Object Inclusion Filtering in Fig. 9, while Section IV-A6 describes it as transmitting \"a spatially sparse feature map\" to reduce raw sensor data, which matches the paper's own Data Format Optimisation definition. Section IV also concedes that MCO \"does not strictly fit the classification,\" and Section IV-D introduces an Infrastructure-Based Redundancy Mitigation Mechanism as a further category not shown in Fig. 9 or the three-part taxonomy. Because assignments appear to depend on editorial judgment beyond the stated definitions, the taxonomy as presented is neither clearly mutually exclusive nor clearly exhaustive. The authors should either revise the category definitions and assignments to align with the text, or explicitly present the taxonomy as a heuristic organizational device with admitted overlaps and residual categories.","section":"IV; Fig. 9; II-E; IV-A6; IV-D"},{"comment":"The paper repeatedly states that none of the newly specified VoI-based measures have been evaluated in the literature, for example in the Introduction and in the opening paragraph of Section III, which says there have been \"no studies investigating\" the angle-based and classification-confidence-based VoI methods, VoI-based frequency and content management, or the use of MCO. No search methodology, databases, inclusion criteria, or cutoff date are provided, so this absolute negative claim is not verifiable. It is also internally weaker in Section VI-B, where the authors write that \"existing research has not evaluated all eight VoI methods,\" which is a different statement. The authors should either document a systematic literature-search protocol that supports the stronger claim, or rephrase to \"to the best of our knowledge\" with a concrete description of the search scope. This matters because the open-problems section builds its motivation in part on the claimed absence of evaluations.","section":"I; III (opening); VI-B"},{"comment":"The parameter-sensitivity discussion for the Frequency-Based RMR refers to \"Fig. 12\" when illustrating the dependencies between N_Redundancy and W_Redundancy, but the relevant figure in the manuscript is Fig. 7; the Dynamics-Based discussion refers to Fig. 8. Since the surrounding analysis explicitly depends on these figures, the cross-reference should be corrected. In addition, Figs. 7 and 8 appear to be the authors' own qualitative summaries rather than results reproduced from Delooz et al.; if so, the text should say so explicitly, since the current wording could be read as reporting the original paper's conclusions.","section":"III-B; Fig. 7; Fig. 8"}],"minor_comments":[{"comment":"The figure caption contains a typo: \"required fror\" should be \"required for\".","section":"Fig. 1"},{"comment":"The acronym table lists \"MTU\" and \"RMR\" twice; one duplicate of each should be removed.","section":"Table I"},{"comment":"The first paragraph of Section IV-D contains a typo: \"redundnacy mitigation\" should be \"redundancy mitigation\".","section":"IV-D"},{"comment":"The sentence \"showing significant reductions in the message sizes transmitted between agents, expressed as a log scale\" is unclear; please specify what is plotted on the logarithmic axis and which dataset or condition the 25.81% AP improvement refers to.","section":"IV-A6"},{"comment":"The sentence \"Authors in [27] are the only to date to have attempted...\" should read \"The authors in [27] are the only ones to date to have attempted...\".","section":"VI-B"},{"comment":"The claim that realistic sensors can only achieve 42% of the object detection rate of idealized sensors, attributed to [46], would benefit from a brief note on the scenario and sensing model used, since it is presented as a general quantitative finding.","section":"V"}],"recommendation":"major_revision","confidential_remarks":"The manuscript is within scope and likely salvageable, but the taxonomic inconsistencies and the overstrong claim about the absence of evaluations should be addressed before publication. I do not see a need for rejection: the survey content is broad and current, and the figures/tables are generally well designed. The main risk is that a reader relying on the three-category taxonomy will find conflicting assignments between Fig. 9 and the body text."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"This survey is worth reading for anyone entering the redundancy mitigation space for V2X collective perception. What is actually new: a three-way taxonomy (object inclusion filtering, data format optimisation, frequency management), a close comparison of the 2019 ETSI TR and 2023 TS with the shift to VoI-based measures, and a reasonably comprehensive review of the academic literature, including simulation environments and open issues. The best part is the detailed walk through the ETSI standards evolution; I know of no other survey that does this as thoroughly.\n\nThe taxonomy is the main value-add, but it is not applied cleanly. The authors themselves concede that MCO does not strictly fit, and they add an Infrastructure-Based category outside the three classes. More concretely, Where2comm is grouped under object inclusion filtering even though their own description says it transmits a spatially sparse feature map, which sounds like data format optimisation. Frequency and Content Management is placed only under frequency management despite jointly selecting content. This fuzziness does not kill the survey, but it does undercut the \"clear framework\" claim. A careful revision should either refine the category definitions or acknowledge that some methods span classes.\n\nThere is also a cross-reference error: Section III-B refers to Fig. 12 for the N_Redundancy/W_Redundancy dependencies, but that figure is actually Fig. 7 (with Fig. 8 for P/S thresholds). And the recurring claim that \"none of the newly specified VoI-based measures have been evaluated in the literature\" is asserted without a documented search protocol. As a load-bearing negative claim for the open problems section, it should be qualified or backed by a systematic search.\n\nI cannot verify all reported numbers from primary sources, but the paper is descriptive and the citations look appropriate. The authors do not hide the limitations of the approaches they review. The survey is a useful reference, not a breakthrough. It deserves a serious referee; the taxonomy issues and the literature-search claim are fixable with revision.\n\nIf you work on CPS or V2X communication, bring it to the reading group. I would cite it as a starting point. My recommendation: send to peer review, but the authors should be asked to tighten the taxonomy definitions and substantiate the \"no evaluations\" claim.","headline":"A genuinely useful survey of ETSI's redundancy mitigation evolution, but the three-way taxonomy is applied loosely enough that the central 'clear framework' claim needs tightening.","tokens_in":35575,"tokens_out":2118,"would_cite":true,"duration_ms":21671,"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 survey proposes a three-category taxonomy—object inclusion filtering, data format optimisation, and frequency management—to organise the fragmented literature on redundancy mitigation for vehicular cooperative perception.","keywords":["Collective Perception Service","redundancy mitigation","Value of Information","ETSI CPM","vehicular networks","cooperative perception","survey taxonomy"],"falsifier":"Locate any peer-reviewed evaluation of the angle-based VoI, classification-confidence-based VoI, or VoI-based frequency/content management measures from ETSI TS 103 324; if such an evaluation exists, the paper's statement that none of the newly specified VoI-based measures have been evaluated in the literature is false.","tokens_in":34621,"feed_emoji":"🚗","tokens_out":5636,"duration_ms":48271,"temperature":0.7,"pith_summary":"This survey argues that the field of redundancy mitigation for the vehicular Collective Perception Service (CPS) is fragmented and needs a common structure. It proposes a taxonomy that divides all mitigation approaches into three classes: object inclusion filtering (which objects to report), data format optimisation (how many bytes per object), and frequency management (how often to send). The paper also traces the evolution of the European standards from the 2019 rules to the 2023 Value of Information (VoI) based specification, and it claims that none of the newly specified VoI measures have been evaluated in the literature. A sympathetic reader would take the paper's aim as giving researchers a shared map for comparing existing schemes and positioning future work.","feed_headline":"A three-way taxonomy organizes redundancy fixes for car-to-car perception","feed_subtitle":"A fragmented field gets one framework that places standards, VoI rules, and research proposals side by side.","key_machinery":"The machinery that carries the argument is the taxonomy itself, together with the VoI concept it organises. A Value of Information (VoI) measure scores each detected object by how much new, useful knowledge transmitting it would give other stations; eight such measures are specified in the 2023 standard. The taxonomy groups mitigation schemes by which resource they conserve: object inclusion filtering removes entire objects from the message, data format optimisation shrinks the data per object, and frequency management sends fewer messages overall. The paper uses this structure to position every surveyed method and to expose what has not been evaluated.","core_discovery":"The paper's central claim is that every redundancy mitigation strategy for CPS can be understood as acting on one of three levers: deciding which detected objects deserve a slot in a Collective Perception Message, reducing the amount of data carried about each included object, or lowering the rate at which messages are generated. On this taxonomy, the standards themselves line up naturally: the ETSI VoI-based inclusion rules belong to object inclusion filtering, the VoI-based frequency and content management belongs to frequency management, and Multi-Channel Operation is acknowledged as an outlier that prioritises channel resources rather than reducing redundancy. The paper further claims that the 2023 standard extended the 2019 rules with two new VoI measures (angle-based and classification confidence-based), combined measures, and a frequency/content management proposal, and that none of the newly specified VoI-based measures have been evaluated in the literature.","pith_inferences":["The paper leaves implicit that its taxonomy is a claim about completeness; a hybrid scheme that simultaneously filters objects and compresses data would need a principled place, and the paper's own example of Multi-Channel Operation shows a fourth category may be needed for resource-prioritisation approaches.","The strongest test of the taxonomy is whether it predicts new work: if the taxonomy is sound, future methods should naturally map onto one category, and proposals that straddle categories should reveal themselves as integration challenges.","Its undocumented 'not evaluated' claim could be checked by an independent literature review, which would either confirm the gap or force a revision of the open-problems section."],"forward_implications":["Researchers can use the taxonomy to position a new mitigation scheme quickly and to compare it against methods from the same category.","The open problems section becomes a roadmap: combined VoI methods, adaptive thresholds, and MCO-based prioritisation are identified as untested and under-specified.","If the taxonomy is complete, the field's fragmentation is a matter of vocabulary rather than substance, and the three categories could become a common language for standards and academic work alike.","The paper's claim that the new VoI measures are unevaluated, if true, means early evaluation studies of those measures would have clear novelty."],"supporting_citations":[{"why":"the 2019 ETSI technical report that first specifies the redundancy mitigation rules and their simulation study.","marker":"[3]"},{"why":"the 2023 ETSI technical specification that extends the rules into VoI-based object inclusion rate control and frequency/content management.","marker":"[4]"},{"why":"the Pre-filter proposal that quantifies CPM generation load and forms the basis for the dynamics-based VoI measure.","marker":"[9]"},{"why":"the Look-Ahead proposal showing default inclusion rules produce many small CPMs, and the basis for a frequency-management alternative.","marker":"[18]"},{"why":"the evaluation of four of the six original RMRs across channel load, perception accuracy, and redundancy level.","marker":"[29]"},{"why":"the earlier survey whose rule/distance/learning taxonomy the paper contrasts with its own conceptual classification.","marker":"[5]"},{"why":"the value-anticipating method that predicts recipient usefulness and links redundancy mitigation to control decisions.","marker":"[27]"}],"fun_headline_variants":["Three levers tame redundant data in cooperative car perception","Value-of-information rules evolve to curb message clutter on the road","Taxonomy of redundancy mitigation for car-to-car collective perception","Open issues for pruning redundant data in vehicular cooperative sensing","Standard's VoI rules evolve to cut redundancy in collective perception"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The paper's claim that none of the newly specified VoI-based measures have been evaluated depends on an undocumented and unverifiable literature search; a single published evaluation of one of those measures would weaken the paper's central justification for those open problems.","fun_headline_variants_meta":{"raw":{"variants":["Three levers tame redundant data in cooperative car perception","Value-of-information rules evolve to curb message clutter on the road","Taxonomy of redundancy mitigation for car-to-car collective perception","Open issues for pruning redundant data in vehicular cooperative sensing","Standard's VoI rules evolve to cut redundancy in collective perception"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000707,"raw_usage":{"total_tokens":3148,"prompt_tokens":867,"completion_tokens":2281,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":483,"completion_tokens_details":{"reasoning_tokens":2198}},"tokens_in":483,"tokens_out":2281,"duration_ms":15813,"temperature":1.0,"reasoning_tokens":2198,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-10T22:32:19.047802+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Locate any peer-reviewed evaluation of the angle-based VoI, classification-confidence-based VoI, or VoI-based frequency/content management measures from ETSI TS 103 324; if such an evaluation exists, the paper's statement that none of the newly specified VoI-based measures have been evaluated in the literature is false.","supporting_citations":[{"cited_title":"Intelligent transport system (its); vehicular communications,","cited_arxiv_id":null,"evidence_quote":"the 2019 ETSI technical report that first specifies the redundancy mitigation rules and their simulation study."},{"cited_title":"Intelligent transport system (its); vehicular communications,","cited_arxiv_id":null,"evidence_quote":"the 2023 ETSI technical specification that extends the rules into VoI-based object inclusion rate control and frequency/content management."},{"cited_title":"Redundancy mitiga- tion in cooperative perception for connected and automated vehicles,","cited_arxiv_id":null,"evidence_quote":"the Pre-filter proposal that quantifies CPM generation load and forms the basis for the dynamics-based VoI measure."},{"cited_title":"Generation of co- operative perception messages for connected and automated vehicles,","cited_arxiv_id":null,"evidence_quote":"the Look-Ahead proposal showing default inclusion rules produce many small CPMs, and the basis for a frequency-management alternative."},{"cited_title":"Analysis and evaluation of information redundancy mitigation for V2X collective perception,","cited_arxiv_id":null,"evidence_quote":"the evaluation of four of the six original RMRs across channel load, perception accuracy, and redundancy level."},{"cited_title":"Value- anticipating V2V communications for cooperative perception,","cited_arxiv_id":null,"evidence_quote":"the value-anticipating method that predicts recipient usefulness and links redundancy mitigation to control decisions."}],"review_version":1}