{"id":"f4dc1840-bdd9-41f3-bde2-a9bc816b012a","arxiv_id":"2412.20699","paper_version":2,"verdict":"CONDITIONAL","confidence":"HIGH","novelty_score":3.0,"correctness_risk":"low","formal_verification":"none","parameter_count":0,"one_line_summary":"A review of UAV-UGV collaborative mapping and navigation for fire and rescue, with a classification of systems by team composition and a discussion of open challenges.","lead":"This paper reviews how aerial drones and ground robots cooperate in fire and rescue missions, focusing on drone-built maps and ground-robot navigation. It categorizes these systems by how many drones and ground robots are involved, then surveys mapping, localization, and path planning methods.","discovery_kind":"review","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The survey's value claim—being a 'systematic review' and a useful reference for fire and rescue mapping/navigation—rests on an undocumented selection process; without one, coverage and the Section V taxonomy cannot be verified.","rationale":"The reader's conditional verdict identifies the same load-bearing assumption: the survey's value depends on comprehensive selection and an apt taxonomy, neither of which is supported by a described methodology. My reading confirms this and does not find a reason to move the verdict. The paper is internally organized and the map/navigation summaries are largely consistent with the literature, so the concern is not about factual errors in individual entries. Instead, the risk is that the paper overstates its status as a systematic review. If the proposed corpus-reconstruction test shows high recall and that all relevant works fit the four categories, the conditional could be lifted; if it shows low recall or poor category fit, the 'systematic' claim should be softened or the taxonomy revised. No change to the reader's conditional verdict is warranted on the current evidence.","tokens_in":27486,"tokens_out":3199,"duration_ms":36003,"concrete_test":"Reconstruct the review corpus with a reproducible search: query IEEE Xplore, Scopus, and Web of Science for ('UAV' OR 'aerial' OR 'drone') AND ('UGV' OR 'ground robot') AND ('fire' OR 'rescue' OR 'disaster') AND ('mapping' OR 'navigation' OR 'localization'), with explicit inclusion criteria and dual screening. Compare the retrieved set against the 151 references in the paper and compute recall of clearly relevant works; then re-apply the Section V four-category assignment to the full retrieved set. If the paper omits a substantial set of relevant studies, or if many retrieved works do not fit the four categories, the 'systematic' and 'comprehensive' descriptors should be removed or the taxonomy revised.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The paper's central claim is that it provides a systematic, comprehensive review of air-ground collaborative mapping and navigation for fire and rescue (Abstract; Section I). That claim is only as strong as the completeness and aptness of the surveyed literature. The paper has no methodology section: no search databases, query strings, date range, inclusion/exclusion criteria, or screening process are reported. The classification in Section V into single-single, single-multi, multi-single, and multi-multi systems is presented as natural, but the paper does not justify why robot count is the organizing dimension for fire-and-rescue mapping/navigation, nor does it show the categories are exhaustive. Several entries in Section V (e.g., Tanner [137] on target detection, Nazarova et al. [136] on earthquake rescue) are not primarily mapping/navigation or fire-rescue studies, so their placement does not directly support practical conclusions. Because the survey introduces no new experimental results, its usefulness to practitioners is exactly its coverage and organization; both are assumed rather than demonstrated.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper is a survey of air-ground collaborative robots for fire and rescue missions, organized around the roles of UAVs in mapping and UGVs in navigation. It introduces a framework in which UAVs build maps (2D grid, 3D, topological, semantic) that UGVs use for co-localization and path planning, and it classifies collaborative systems into four categories based on the number of UAVs and UGVs: single-single, single-multi, multi-single, and multi-multi. It then presents application examples in firefighting and rescue scenarios and closes with research challenges and future directions. The paper does not report new experimental results; its contribution is a structured synthesis and classification of existing work.","tokens_in":27676,"tokens_out":3955,"duration_ms":40580,"significance":"If the survey's coverage and taxonomy are sound, the paper would provide a useful entry point for practitioners and researchers in a fragmented area, and its tables summarising mapping types, localisation methods, and team configurations are a helpful reference resource. The authors also credit a specific application-oriented perspective (UAV mapping for UGV navigation) that distinguishes the survey from earlier reviews. However, the value of the survey depends on the completeness of the literature selection and on whether the proposed classification is apt; both are assumed rather than demonstrated. The paper is therefore a potentially valuable reference, but its central claim to being a systematic review is not yet supported by the manuscript as written.","major_comments":[{"comment":"The paper claims to present a 'systematic review' of air-ground collaborative robots for fire and rescue from a mapping and navigation perspective, but it provides no methodology section: no search databases, query strings, date range, inclusion/exclusion criteria, or screening process are reported. Without such a protocol, the reader cannot verify that the surveyed literature is comprehensive or representative, which is load-bearing because the paper introduces no new experimental results and its usefulness rests entirely on its coverage and organization.","section":"Section I (Introduction, overall structure)"},{"comment":"The classification of collaborative systems into four categories based on the number of UAVs and UGVs is presented as natural without justification. The paper does not explain why robot count is the most meaningful organizing dimension for fire-and-rescue mapping and navigation, or why other dimensions such as communication architecture, autonomy level, or map-sharing strategy are less important. Several entries placed in the taxonomy do not directly support the mapping/navigation focus: Tanner [137] concerns target detection and Nazarova et al. [136] addresses earthquake rescue using search theory. Such placements weaken the paper's practical conclusions about mapping and navigation and suggest the taxonomy is not consistently applied.","section":"Section V and Table IV"},{"comment":"The tables of 'representative works' do not state how the entries were selected or how many papers were screened to arrive at them. As a result, the reader cannot determine whether the tables are balanced or biased, and the claim that the paper 'rounds up references for practitioners' is unsubstantiated. The paper also gives generic pros and cons for map types and navigation methods without tying these assessments to the experimental conditions or performance metrics of the cited works; for example, the merits and demerits of lidar localization (Section IV-A.4) are stated in general terms rather than derived from the reviewed studies. A critical synthesis that explains which methods transfer to fire/rescue scenarios would strengthen the survey.","section":"Tables I, II, and III"}],"minor_comments":[{"comment":"The phrase 'ground-to-ground cooperative robots' in the abstract contradicts the title and the rest of the paper, which is about air-ground collaboration; this should be corrected to 'air-ground'.","section":"Abstract"},{"comment":"The sentence 'which is the basisn' appears at the end of the section and is missing a period and a word; it should read 'which is the basis' or similar.","section":"Section I"},{"comment":"The phrase 'the UAG' should be 'the UAV'.","section":"Section III-A"},{"comment":"In the row for Zuo et al. [105], 'V oronoi' should be 'Voronoi'.","section":"Section IV-B and Table III"},{"comment":"Spacing between 'UA' and 'V' (e.g., 'UA V', 'UGV ') is likely a LaTeX rendering artifact, but the camera-ready version should ensure the abbreviations appear as 'UAV' and 'UGV' consistently.","section":"Throughout"},{"comment":"Reference [63] contains '2Proc' at the start of the venue name, which appears to be a typographical error.","section":"References"}],"recommendation":"major_revision","confidential_remarks":"The paper is a survey and its contribution is organizational rather than experimental. The main concerns are the absence of a literature-selection protocol and the lack of a defensible basis for the Section V taxonomy; both are fixable in revision. The authors also cite a number of their own prior works (e.g., [7], [60], [102], [103], [141]) in support of key claims; while self-citation is not improper, the survey would be strengthened by broader independent validation and by an explicit acknowledgment of this pattern in the text."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Bottom line: this is a serviceable entry-level survey of UAV-UGV mapping and navigation for fire and rescue, with one honest organizing idea and a load-bearing overclaim. The four-way split by robot count is not new, but tying it to map type and navigation method in a single framework is a useful prism. The authors give a decent structured tour of grid, 3D, topological, and semantic maps, and they are candid about each type's limitations. The tables are a genuine asset, and the separation of mapping from co-localization from UGV navigation is clean.\n\nThe soft spot is the word 'systematic.' There is no methodology section: no database list, query strings, dates, or inclusion/exclusion criteria, so the coverage and the Section V taxonomy cannot be verified. The taxonomy itself is presented as natural but the paper never argues why robot count is the axis that matters for fire and rescue missions; some cited entries, like Tanner's target-detection work and Nazarova's earthquake-rescue study, are not primarily mapping/navigation or fire-rescue studies. The writing also has repeated typos ('ground-to-ground' in the abstract, 'the UAG', 'V oronoi'), and the abstract's phrase 'ground-to-ground cooperative robots' is just wrong for an air-ground paper. These are fixable but unprofessional. The self-citations are numerous but not egregious; they support the framework rather than carrying it.\n\nThe survey's central value as a reference depends on selection judgment, and the paper doesn't show that judgment. That said, for a reader new to the area this is a legitimate first stop. For specialists it maps the terrain rather than moving it. It deserves a serious referee, but I'd require the authors to either add a short methodology paragraph or soften 'systematic' to 'structured,' and to justify the four-category taxonomy. I'd also have a copy editor catch the typos before accepting.","headline":"A useful entry-level survey of UAV-UGV mapping and navigation for fire and rescue, but the 'systematic' label is unsupported by any documented method and the taxonomy needs defense.","tokens_in":28243,"tokens_out":1652,"would_cite":false,"duration_ms":17301,"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":"Mapping and navigation decide air-ground rescue robot success","keywords":["air-ground collaborative robots","fire and rescue","UAV mapping","UGV navigation","path planning","co-localization","topological map","semantic map"],"falsifier":"Run a reproducible database search for peer-reviewed UAV-UGV fire-and-rescue mapping and navigation papers from 2015 to 2025 with explicit inclusion criteria; if a substantial share of the recovered papers (say, more than 20%) fits none of the four map types or four team-size categories, the survey's systematic-coverage claim fails.","tokens_in":27281,"feed_emoji":"🚒","tokens_out":7434,"duration_ms":63658,"temperature":0.7,"pith_summary":"This paper claims that the efficiency of air-ground collaborative robots in fire and rescue missions rests on a two-stage pipeline: an unmanned aerial vehicle (UAV) maps the scene from above, and an unmanned ground vehicle (UGV) uses that map to localize, plan, and navigate. It argues that no prior survey has examined this domain through the mapping-and-navigation lens, and it delivers a systematic review organized by two axes: the type of map the UAV builds (2-D grid, 3-D, topological, semantic) and the number of UAVs and UGVs in the team (single-single, single-multi, multi-single, multi-multi). The survey catalogs representative works, their merits and demerits, and matches navigation algorithms to map types. If the paper is right, practitioners gain a structured reference for choosing mapping and navigation methods for specific fire-and-rescue deployments, and researchers gain a map of open problems.","feed_headline":"Mapping and navigation decide air-ground rescue robot success","feed_subtitle":"A systematic review organizes UAV/UGV fire-and-rescue systems by map type and team size, with open challenges.","key_machinery":"The central organizing device is the UAV-mapping-to-UGV-navigation pipeline: the UAV builds a map (2-D grid, 3-D point cloud or octree, topological graph, or semantic map), the UGV co-localizes within that map, and then plans and executes navigation. The second load-bearing device is the team-cardinality classification (1 UAV–1 UGV, 1 UAV–many UGVs, many UAVs–1 UGV, many UAVs–many UGVs), which the paper uses to discuss scalability, communication burden, and mission practicality. Together these two axes carry the survey's argument that the field is coherent and that its open challenges are identifiable.","core_discovery":"The central claim is that the design space of air-ground collaborative robots for fire and rescue can be organized by a clear division of labor: UAVs, with their mobility and aerial sensors, are responsible for rapid large-scale perception and map construction; UGVs, with their ground-level access, are responsible for co-localization, path planning, and navigation toward mission points. The paper asserts that this mapping-then-navigation framework is the key foundation for efficient autonomous collaboration, and it substantiates the claim by classifying maps into four types, localization into five sensor strategies, navigation algorithms by map type, and teams by cardinality of UAVs and UGVs. The survey's contribution is taxonomic and organizational: it provides a coherent reference that situates existing results, highlights their limitations, and identifies research gaps such as task-relevant mapping, multi-modal data fusion in smoke-filled environments, and embodied-AI-driven navigation.","pith_inferences":["One implicit consequence is that the map type and team cardinality axes are interdependent: for example, semantic maps look more attractive in many-many systems where the cost of building them can be amortized across multiple ground vehicles, a link the paper does not explicitly make.","A testable extension is to turn the survey's taxonomy into a benchmark: define representative fire-and-rescue scenarios (indoor high-rise, forest wildfire, chemical plant) and score how well each map-type and team-size combination performs, which would validate or refine the paper's organizational claims.","The framework suggests that the communication channel between UAV and UGV—bandwidth, latency, reliability—is a hidden variable that may dominate the choice of map type; the paper mentions communication challenges but does not elevate channel constraints to a first-class design axis alongside map type and team size."],"forward_implications":["A practitioner choosing a fire-and-rescue system can use the four map types as a decision menu: 2-D grid maps for simple, compute-limited settings; 3-D maps for rugged terrain; topological maps for structured road networks; semantic maps for scene understanding and high-level task planning.","Matching navigation algorithms to map types becomes systematic: grid-based planners like A* and D* for 2-D maps, sampling-based planners like RRT on point clouds for 3-D maps, graph search on topological maps, and semantic-informed planners for semantic maps.","The team-size taxonomy predicts capability trade-offs: single-single systems are simple but limited in scale; single-UAV multi-UGV systems depend on a single aerial vantage and risk a bottleneck; multi-UAV single-UGV systems provide coverage but limit ground mission throughput; many-many systems maximize capability but face coordination, communication, and task-allocation challenges.","For researchers, the survey's identified gaps—task-oriented mapping, multi-modal fusion under smoke, and embodied-AI navigation—are concrete openings for new work."],"supporting_citations":[{"why":"Supplies the prior review baseline of collaborative air-ground robots that this survey extends with a mapping-and-navigation lens.","marker":"[12]"},{"why":"Provides an existing taxonomy and optimization framework for UAV–UGV coordination that this paper positions against.","marker":"[18]"},{"why":"Discusses synchronous planning and control of aerial-ground systems, giving the integration context for the proposed pipeline.","marker":"[19]"},{"why":"Reviews the air-ground collaborative system and identifies four functional roles, a prior organizational scheme the paper refines.","marker":"[20]"},{"why":"Demonstrates UAV-built occupancy grid maps used for UGV path planning, a key example of the 2-D grid map category.","marker":"[34]"},{"why":"Shows a UAV constructing a 3-D OctoMap to guide UGVs, supporting the 3-D map category and the mapping-to-navigation pipeline.","marker":"[38]"},{"why":"Provides a real-time semantic map from a UAV enabling multi-UGV localization and navigation, anchoring the semantic map and single-UAV-multi-UGV categories.","marker":"[45]"},{"why":"Presents a multi-UAV mapping and surveillance system for emergency response, supporting the multi-UAV team category and application examples.","marker":"[124]"},{"why":"Uses a multi-UAV multi-UGV team for forest fire detection and monitoring, grounding the many-to-many configuration.","marker":"[128]"}],"fun_headline_variants":["UAVs map, UGVs navigate: fire rescue robot taxonomy","Systematic review maps air-ground rescue robot design space","Fire rescue robots: UAV mapping, UGV navigation framework","How UAV-UGV teams map and navigate for fire rescue"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The survey's value depends on its selection of representative works being comprehensive and its two classification axes (map type and team size) being the most meaningful way to organize the field; if major methods or system configurations fall outside these categories, the organizational claim weakens.","fun_headline_variants_meta":{"raw":{"variants":["UAVs map, UGVs navigate: fire rescue robot taxonomy","Systematic review maps air-ground rescue robot design space","Fire rescue robots: UAV mapping, UGV navigation framework","How UAV-UGV teams map and navigate for fire rescue"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000407,"raw_usage":{"total_tokens":2141,"prompt_tokens":997,"completion_tokens":1144,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":613,"completion_tokens_details":{"reasoning_tokens":1075}},"tokens_in":613,"tokens_out":1144,"duration_ms":10799,"temperature":1.0,"reasoning_tokens":1075,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-10T23:13:22.563098+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Run a reproducible database search for peer-reviewed UAV-UGV fire-and-rescue mapping and navigation papers from 2015 to 2025 with explicit inclusion criteria; if a substantial share of the recovered papers (say, more than 20%) fits none of the four map types or four team-size categories, the survey's systematic-coverage claim fails.","supporting_citations":[{"cited_title":"A scenario for a multi-uav mapping and surveillance system in emergency response applications,","cited_arxiv_id":null,"evidence_quote":"Presents a multi-UAV mapping and surveillance system for emergency response, supporting the multi-UAV team category and application examples."},{"cited_title":"Cooperative forest monitoring and fire detection using a team of uavs-ugvs,","cited_arxiv_id":null,"evidence_quote":"Uses a multi-UAV multi-UGV team for forest fire detection and monitoring, grounding the many-to-many configuration."}],"review_version":1}