{"id":"23d372ac-8e73-4c5a-9738-521e4d720be2","arxiv_id":"2506.02217","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":4,"one_line_summary":"The paper presents EMMS, a tool for importing real public transportation data into SUMO simulations, and uses it to show that bus contact patterns are short-lived and sparse, favoring opportunistic and delay-tolerant networking.","lead":"A new tool called EMMS automatically converts real bus routes, stops, and timetables from Curitiba, Brazil into SUMO traffic simulations. The simulations show that bus-to-bus contacts are short (around 50 seconds) and separated by longer gaps, suggesting opportunistic data networks with small temporary groups.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Travel-time compatibility checks (Section IV) validate aggregate journey durations but not the fine-grained relative bus separation that determines contact and inter-contact times; Figures 8 and 9 therefore carry an unvalidated simulation-fidelity assumption.","rationale":"The reader's weakest_assumption identifies the same load-bearing gap: travel-time compatibility does not constrain contact-level spatial coupling. I agree with that assessment, so no verdict change is needed. The concern can be sharpened: the compatibility numbers (95.2%, 89.2%, 90.5%, 91.6% in Section IV) are aggregate trip-duration overlaps, while contact time is a joint property of two buses' relative trajectories. Even if each bus's journey time is marginally correct, correlated errors in speed profiles, signal phases, and dwell times could easily shift relative offsets by tens of seconds, which is the same order as the reported mean contact times. The paper provides no sensitivity analysis or direct ground-truth comparison for contacts. The qualitative conclusion—bus contacts are intermittent and services should be delay-tolerant—is robust and likely correct for an urban bus network, but the specific means quoted in the abstract and conclusion are only as reliable as the unvalidated simulation. The direct re-analysis of URBS GPS data would settle the issue without new field experiments. Therefore the appropriate disposition is unchanged: conditional acceptance, with simulation-to-reality contact validation as a requirement.","tokens_in":10455,"tokens_out":2738,"duration_ms":25742,"concrete_test":"Run the same contact/inter-contact analysis directly on the URBS geolocation traces used in the paper (October 25, 2022, 16:00–20:00): compute inter-bus distances from the raw coordinates, apply the 150 m and 300 m thresholds and 3 s sampling exactly as in Section V, and compare the resulting contact-time and inter-contact-time distributions with Figures 8 and 9. If the real means differ from the simulated means by more than ~30%, or if the inter-contact-time/contact-time ratio differs substantially from 2–3, the simulated contact statistics are not validated and the central claim is unsupported.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central quantitative claim—mean contact times of 40–107 s and inter-contact times of 132–163 s at 150/300 m ranges (Section V, Figures 8–9)—depends on the simulated relative positions of buses being realistic. The only validation offered (Section IV) is a ~90% compatibility between real and simulated travel times. Journey time is an aggregate, origin-to-destination quantity; it is insensitive to the small speed variations, traffic-signal phase offsets, dwell times at stops, and bus bunching that determine whether two specific buses come within 150–300 m of each other at the same moment. The paper itself asserts without evidence that high travel-time compatibility 'ensures that the results obtained from the simulation environment are reliable.' No real contact or inter-vehicle-distance measurements are compared against the simulated contact statistics. The 3 s sampling period adds a further quantisation that could truncate very short contacts, but the dominant gap is the missing validation of the spatial-temporal coupling. Because the conclusion that communication services must be opportunistic and delay-tolerant is keyed to these specific short-contact, long-inter-contact statistics, this unvalidated fidelity assumption is load-bearing.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper presents EMMS, a tool that automatically converts real public transportation data (routes, bus stop locations, and departure timetables) into SUMO simulation configuration files via map-matching. The authors use EMMS to simulate bus mobility in two regions of Curitiba (central and southern) using data from October 25, 2022, and then report a statistical analysis of travel times, vehicle density, connectivity, and contact/inter-contact times. The central claim is that bus-to-bus contacts are short (mean contact times of 57-107 s for 150-300 m transmission ranges), inter-contact times are about two to three times longer (132-163 s), and the resulting groups are small (2-3 buses), so any communication service built on this network must be opportunistic and delay-tolerant.","tokens_in":10724,"tokens_out":4762,"duration_ms":42571,"significance":"The EMMS pipeline is a practical engineering contribution: it automates the labor-intensive process of converting real bus routes, stops, and timetables into a SUMO simulation, and the paper reports real-data-driven mobility statistics for two distinct urban regions. The travel-time compatibility check (~90% in Section IV) is a useful sanity check that the simulated routes and schedules broadly match reality. If the fine-grained contact statistics were validated, the paper would provide valuable input for the design of opportunistic and delay-tolerant communication protocols in public-transportation-based networks. However, the headline quantitative claim rests on an unvalidated fidelity assumption, as detailed in the major comments.","major_comments":[{"comment":"The validation offered for the simulation is limited to travel-time compatibility (~90% between real and simulated journey times). Travel time is an aggregate, origin-to-destination quantity; it is not sensitive to the small speed variations, traffic-signal phase offsets, dwell times, and bus-bunching effects that determine whether two buses come within 150-300 m of each other at the same moment. The sentence that a high degree of compatibility 'ensures that the results obtained from the simulation environment are reliable' is therefore an overreach with respect to the contact/inter-contact statistics in Figures 8 and 9. Since the paper's central conclusion about opportunistic and delay-tolerant communication is keyed to these statistics, this unvalidated simulation-fidelity assumption is load-bearing. The authors should compare simulated contact durations and inter-bus distances against real GPS-derived measurements from the URBS dataset, or, if that is not feasible, explicitly limit the claim to the simulation environment and describe the contact statistics as not independently validated.","section":"§IV (validation paragraph) and §V (Figs. 8-9)"},{"comment":"The reported means (e.g., 57 s contact time and 132 s inter-contact time at 150 m in the central region; 107 s and 163 s at 300 m) are presented without confidence intervals, standard deviations, or sample sizes. The violin plots show distribution shapes, but the text makes precise mean-based claims without any measure of uncertainty. The authors should report the number of contact events, the dispersion around each mean, and, where comparisons are made (between regions or between ranges), appropriate statistical tests or at least non-overlapping confidence intervals. Without these, the reader cannot judge whether the differences described are meaningful.","section":"§V, 'Contact Time and Inter-contact Time' (Figs. 8-9)"},{"comment":"The paper states that 'data sampling was performed every 3 seconds.' This sampling period imposes a lower bound on the duration of a detectable contact and quantizes inter-contact times: any contact shorter than 3 s is missed, and the measured contact durations are multiples of the sampling period. The mean contact times reported (57-107 s) are well above 3 s, so the bias may be modest, but the paper provides no sensitivity analysis. The authors should test whether the qualitative conclusion (short contacts, long inter-contact times, groups of 2-3 buses) is robust to the sampling interval, for example by repeating the analysis at 1 s and 5 s sampling periods.","section":"§V, sampling interval"},{"comment":"The simulation appears to include only buses; there is no explicit mention of general-purpose traffic, passenger-induced dwell-time variability, or varying traffic-signal phase offsets, all of which are known to affect bus bunching and the instantaneous relative positions of buses. The travel-time compatibility at the aggregate level does not directly validate these fine-grained dynamics. The authors should either incorporate and validate these elements (e.g., by adding background traffic to the SUMO scenario) or explicitly list them as limitations of the contact and inter-contact statistics. This limitation is especially important because the central claim concerns the exact distribution of short contacts and inter-contact gaps.","section":"§IV, simulation scenario"}],"minor_comments":[{"comment":"The compatibility values in the Conclusion (91.17% central, 90.42% southern) do not match the values in Section IV (afternoon 95.2%/89.2%; morning 90.5%/91.6%). Please reconcile the two sets of numbers or explain how the averages were computed.","section":"§VI vs. §IV"},{"comment":"The abstract and introduction describe the work as a 'complete statistical analysis,' but the paper reports only descriptive statistics and no hypothesis tests or uncertainty quantification. Please rephrase to avoid overclaiming.","section":"Abstract and §I"},{"comment":"The definition of inter-contact time in Figure 3 is confusing: the caption says 't4 being the beginning of the connected interval' for bus B, which appears to contradict the main text definition. Clarify with a precise timeline annotation.","section":"Figure 3"},{"comment":"The 150 m and 300 m transmission ranges are cited to reference [29], which is an IEEE 802.11p performance study; the choice should be justified in one sentence (e.g., as representative of V2V communication ranges) rather than simply cited.","section":"§V, transmission range choice"},{"comment":"The paper does not state the weekday/weekend status, weather conditions, or any special events on October 25, 2022, which could affect traffic and bus behavior. Adding these contextual details would help readers assess the generalizability of the results.","section":"§IV, data context"},{"comment":"The conclusion mentions 'results obtained for other dates and time periods (omitted here due to space limitations)' without any supporting data. Either include a representative figure or state clearly that the paper analyzes only the single date and period described in Section IV.","section":"§VI, omitted results"}],"recommendation":"major_revision","confidential_remarks":"The paper fits the scope of the journal and the EMMS tool is a solid practical contribution. The main gate for acceptance is the unvalidated fidelity of the contact dynamics: the travel-time check is not sufficient to support the central quantitative claim. The authors should be required to add either a real-data contact validation (using the available URBS GPS data) or, at a minimum, a clear and prominent limitation statement and a sensitivity analysis of the sampling interval. The inconsistency in the compatibility percentages between Section IV and Section VI should also be fixed."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The paper is worth a look for two reasons: it presents EMMS, a real tool that automates map-matching of bus routes, stops, and timetables into SUMO, and it reports new contact-pattern statistics for Curitiba's bus network. The travel-time validation against URBS data (89–95% compatibility) is a legitimate check for aggregate journey realism, and the violin plots for travel time, density, connectivity, and contact/inter-contact time give a reasonable descriptive picture of a bus-based opportunistic network.\n\nThe central soft spot is exactly what the stress-test note flags: travel-time compatibility does not validate the fine-grained relative positions of buses, and those relative positions determine the contact and inter-contact times that form the core results. Two buses can have realistic origin-to-destination times while the moments they come within 150–300 m of each other are wrong by tens of seconds. The paper's claim that high travel-time compatibility 'ensures that the results obtained from the simulation environment are reliable' overreaches. The contact statistics should be presented as simulation outputs whose fidelity is plausible but unproven, unless the authors add a comparison against real GPS-derived contact intervals or inter-bus distance distributions. The 3-second sampling period adds a further caveat: very short contacts are truncated, so the reported means may be slightly biased upward.\n\nOther soft spots are real but minor. The quantitative results come from one afternoon in one city; the authors mention other periods are omitted, which is fine for an exploratory study but limits generality. There are no error bars or statistical tests on the contact/inter-contact means, only descriptive violin plots. The tool and data are not publicly released, which makes reproduction harder. The self-citation [15] is not a problem: it is a prior paper by the same group, not evidence of a circular argument.\n\nThe central qualitative conclusion—that bus contacts are short, inter-contacts are two to three times longer, and groups are small (2–3 buses), so communication services must be opportunistic and delay-tolerant—is plausible and consistent with intuition. The specific numbers (57–107 s contact, 132–163 s inter-contact) should be treated as preliminary case-study outputs rather than general results.\n\nWho is this for? Researchers working on DTN routing or V2V communications using public transit backbones, especially those considering SUMO-based simulation. It is not a methodological breakthrough, but the EMMS contribution and Curitiba dataset are useful. A serious referee should see it, with a request for additional validation of contact dynamics, statistical treatment of the reported means, and ideally release of the tool and data.","headline":"A useful engineering case study with a real tool and new Curitiba contact statistics, but the headline numbers rest on a simulation-fidelity assumption that is only weakly validated by aggregate travel-time checks.","tokens_in":11215,"tokens_out":1220,"would_cite":false,"duration_ms":13082,"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":"Bus-to-bus wireless contacts in a simulated Curitiba transit system last 57-107 seconds, with gaps of 132-163 seconds, implying that any communication service must be opportunistic and delay-tolerant.","keywords":["public transportation systems","vehicular ad hoc networks","opportunistic communication","delay-tolerant networks","contact time","inter-contact time","map-matching","SUMO simulation"],"falsifier":"Compute pairwise distances directly from the original Curitiba GPS traces used as input for the same buses and time window, then derive empirical contact and inter-contact times with the same 150 m and 300 m thresholds and 3-second sampling; if those empirical means differ from the simulated 40-107 second contact times and 132-163 second inter-contact times by more than a few tens of seconds, the travel-time validation is insufficient to support the contact conclusions.","tokens_in":10291,"feed_emoji":"🚌","tokens_out":6696,"duration_ms":59190,"temperature":0.7,"pith_summary":"This paper sets out to measure what wireless contact between buses in a public transportation system actually looks like, using a simulation pipeline that converts the city's real geolocation and timetable data into bus mobility. The authors built a tool, EMMS, that automatically matches bus routes and stops onto the SUMO road network, and they check it by comparing simulated and real travel times. Analyzing 152 buses in two Curitiba regions, they find mean contact times of 57-107 seconds at 150-300 m transmission ranges, and mean inter-contact times of 132-163 seconds, so buses form only small, temporary groups of two or three vehicles. The result matters because it tells network designers that bus-to-bus communication cannot rely on stable links or multihop paths, and must instead be opportunistic and delay-tolerant.","feed_headline":"Bus contacts last under 2 minutes; gaps run 2-3x longer","feed_subtitle":"Simulated Curitiba transit shows small 2-3 bus groups, pointing to opportunistic, delay-tolerant networking.","key_machinery":"The central object is EMMS (Engine for Map-Matching to SUMO), a four-layer Python tool that turns public bus data into a SUMO simulation: layer one validates and converts GPS coordinates; layer two matches each itinerary point to a road edge, enforcing physical and logical edge interconnections; layer three places bus stops on matched edges; layer four writes the configuration files. The analytical machinery is the pair of metrics 'contact time' and 'inter-contact time,' computed from pairwise distances with transmission ranges of 150 m and 300 m, sampled every 3 seconds and summarized as violin plots. The travel-time compatibility check (around 90% agreement between simulated and real bus journey times) is what licenses the claim that these simulated contacts are reliable.","core_discovery":"On the paper's own terms, the contact and inter-contact statistics of the simulated Curitiba public transportation system show a fragmented network. With a 150 m transmission range, mean contact time is about 57 seconds in the central region and 40 seconds in the southern region; with 300 m, it rises to roughly 107 seconds, while inter-contact times stay in the 132-163 second range. Combined with the per-vehicle connectivity results, this means each bus typically has one or two neighbors during a contact and groups average 2-3 buses. The paper takes these numbers as evidence that any communication service running on the bus network would be opportunistic and delay-tolerant by nature.","pith_inferences":["If the same contact statistics hold in other cities, routing parameters for delay-tolerant bus networks could be pre-tuned from public timetable and route data alone, skipping per-city simulation.","Feeding the simulated traces into a packet-level network simulator would quantify how much data actually transfers within a 57-107 second contact at realistic 802.11p rates, which the paper leaves for future work.","The small group sizes suggest buses might serve better as message ferries between fixed access points than as a mesh backbone; this design direction is implied but not tested here.","A testable extension is to repeat the measurement with different traffic-light timings and congestion levels to see whether inter-contact times remain stable, since those are exactly the fine-grained variables the travel-time validation may not constrain."],"forward_implications":["At measured contact durations of 40-107 seconds, an encounter can carry only a limited data volume at typical vehicular communication rates, so applications should be designed for short bursty exchanges.","Inter-contact times of 132-163 seconds impose a minimum delay of minutes between forwarding opportunities, ruling out real-time or streaming services on the bus network alone.","With groups of 2-3 buses per contact, multi-hop paths are rare and fragile; routing should assume frequent disconnections and use store-carry-forward.","Because the two Curitiba regions show similar contact dynamics, the qualitative conclusion of an opportunistic, delay-tolerant network appears robust across different terminal layouts and bus densities.","Beyond these two regions, the paper reports that other dates and periods yield distributions of the same shape, suggesting the design guidance is not limited to the single afternoon analyzed in detail."],"supporting_citations":[{"why":"Supplies the SUMO simulator and the SUMOLib libraries through which EMMS imports routes and stops; the entire trace generation depends on it.","marker":"[14]"},{"why":"Provides the map-matching algorithm taxonomy and techniques that EMMS's edge-matching steps build on.","marker":"[22]"},{"why":"Supplies the candidate-edge region-of-interest approach EMMS uses to find the edges near each GPS point.","marker":"[26]"},{"why":"Provides the real Curitiba bus geolocation, timetables, and travel times that serve as simulation input and as the validation baseline.","marker":"[27]"},{"why":"Gives the IEEE 802.11p transmission range values (150 m and 300 m) used to define when two buses count as in contact.","marker":"[29]"}],"fun_headline_variants":["Bus contacts under 2 min, gaps 2-3x longer","2-3 bus groups, 40-107s contacts in Curitiba transit","Fragmented bus network: opportunistic delay-tolerant links","Short bus contacts, long gaps: opportunistic comms","Transit bus contacts: brief, sparse—ideal for DTN"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The findings hinge on the assumption that the simulated bus trajectories preserve fine-grained spacing and timing between buses, not just average travel times; the paper's roughly 90% travel-time compatibility check does not by itself guarantee that contact-level statistics are realistic.","fun_headline_variants_meta":{"raw":{"variants":["Bus contacts under 2 min, gaps 2-3x longer","2-3 bus groups, 40-107s contacts in Curitiba transit","Fragmented bus network: opportunistic delay-tolerant links","Short bus contacts, long gaps: opportunistic comms","Transit bus contacts: brief, sparse—ideal for DTN"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.00017,"raw_usage":{"total_tokens":1191,"prompt_tokens":794,"completion_tokens":397,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":410,"completion_tokens_details":{"reasoning_tokens":307}},"tokens_in":410,"tokens_out":397,"duration_ms":4295,"temperature":1.0,"reasoning_tokens":307,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-07T11:27:47.197942+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Compute pairwise distances directly from the original Curitiba GPS traces used as input for the same buses and time window, then derive empirical contact and inter-contact times with the same 150 m and 300 m thresholds and 3-second sampling; if those empirical means differ from the simulated 40-107 second contact times and 132-163 second inter-contact times by more than a few tens of seconds, the travel-time validation is insufficient to support the contact conclusions.","supporting_citations":[{"cited_title":"Microscopic traffic simulation using SUMO,","cited_arxiv_id":null,"evidence_quote":"Supplies the SUMO simulator and the SUMOLib libraries through which EMMS imports routes and stops; the entire trace generation depends on it."},{"cited_title":"Current map- matching algorithms for transport applications: state-of-the art and future research directions","cited_arxiv_id":null,"evidence_quote":"Provides the map-matching algorithm taxonomy and techniques that EMMS's edge-matching steps build on."},{"cited_title":"A critical review of real-time map- matching algorithms: Current issues and future directions,","cited_arxiv_id":null,"evidence_quote":"Supplies the candidate-edge region-of-interest approach EMMS uses to find the edges near each GPS point."},{"cited_title":"IEEE 802.11p Performance Evaluation: Simulations vs. Real Experiments,","cited_arxiv_id":null,"evidence_quote":"Gives the IEEE 802.11p transmission range values (150 m and 300 m) used to define when two buses count as in contact."}],"review_version":1}