{"id":"1933b6cc-78ee-4248-b9dd-cce5ce11cf17","arxiv_id":"2501.12792","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":4.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":1,"one_line_summary":"Simulations show that a 5G-TSN network using 3GPP indoor factory channel models can keep control traffic delays low in a single-robot factory scenario.","lead":"This paper uses computer simulation to test a factory wireless network that combines 5G with time-sensitive networking, carrying robot control, video, and background data. The authors report the hybrid network can handle latency-sensitive factory traffic in the simulated indoor environments, but the results come from a simplified setup.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Numerology index 4 at 5.9 GHz is not a valid 3GPP NR configuration; reported latencies may be artifacts of an unsupported subcarrier spacing.","rationale":"The reader's weakest_assumption correctly flagged the numerology choice and the single-UE/FlatGround environment as questionable, but grouped them with broader representativeness concerns. My stress-test sharpens this into a specific, load-bearing technical flaw: numerology index 4 at 5.9 GHz is outside the 3GPP NR specification for FR1, and because latency is the central metric, this configuration directly undermines the quantitative evidence. This is a correctness risk, not a novelty or scope issue. The concern is addressable (choose a valid numerology), so the CONDITIONAL verdict remains appropriate; the paper should be accepted only after the simulations are rerun with a standard configuration and the results are shown to still support the claim. I agree with the reader's overall assessment, and the concrete test would settle whether the headline survives.","tokens_in":9058,"tokens_out":4576,"duration_ms":46623,"concrete_test":"Re-run the simulations of Section IV (Figs. 4 and 6) with the same traffic and channel settings but with a standard FR1 numerology for 5.9 GHz, e.g., numerology 1 (30 kHz SCS) or 2 (60 kHz SCS), replacing numerology index 4. If the end-to-end latencies for NC traffic increase by more than 25% (or cross the latency thresholds of the control applications), the headline conclusion is an artifact of a non-standard configuration. Alternatively, confirm from 3GPP TS 38.104 that numerology 4 is invalid for 5.9 GHz; if so, the reported delays cannot be considered representative of real 5G-TSN deployments.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim that 5G-TSN can address latency-sensitive indoor factory scenarios rests on the simulated end-to-end delays in Figs. 4 and 6. Table I specifies carrier frequency 5.9 GHz (FR1) with numerology index 4. In 3GPP NR, numerology index 4 corresponds to 240 kHz subcarrier spacing (2^4 × 15 kHz), which is not supported in FR1 (410 MHz–7.125 GHz); FR1 deployments use 15, 30, or 60 kHz (indices 0–2), with at most 120 kHz in specific cases. This invalid combination shortens the OFDM symbol duration and slot length (e.g., 0.0625 ms for 240 kHz vs. 0.5 ms for 30 kHz), proportionally reducing physical-layer transmission time, HARQ round-trip time, and processing delays. Since the paper's headline metric is latency, the quantitative values are likely unrealistically optimistic relative to any deployable 5G system in a factory. This is not a matter of consensus but of internal consistency with the standards the paper itself invokes (3GPP TR 38.901/TS 38.211). A second, related inconsistency is the 'FlatGround' physical environment in Table I, which is not representative of an indoor factory and may interact with the InF path-loss implementation. Unless the numerology is corrected and the simulation rerun, the evidence does not support the general claim.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper simulates a 5G-TSN network supporting an automated guided vehicle in an indoor factory, using OMNeT++/Simu5G/5GTQ and implementing the 3GPP TR 38.901 indoor factory (InF) channel profiles. It evaluates SINR, HARQ error rate, and end-to-end delay for three traffic classes (network control, video, best effort) across four InF profiles and three distance ranges, and concludes that 5G-TSN can address latency-sensitive scenarios in indoor factory environments.","tokens_in":9303,"tokens_out":3420,"duration_ms":34999,"significance":"If the quantitative results are valid, the paper would provide a useful early simulation baseline for 5G-TSN deployment in factory environments, and the InF-profile implementation in Simu5G is a concrete contribution. The three-class traffic model and the comparison across InF-SL, InF-DL, InF-SH, and InF-DH are sensible. However, the main evidence rests on a radio configuration that does not correspond to a valid 3GPP NR deployment, so the significance of the reported latency numbers cannot be assessed until the simulations are corrected and rerun.","major_comments":[{"comment":"Numerology index 4 at a 5.9 GHz carrier is not a valid 3GPP NR configuration. Numerology index 4 corresponds to 240 kHz subcarrier spacing, which is defined for FR2, not for FR1 (410 MHz–7.125 GHz), where the supported spacings are 15, 30, and 60 kHz (and at most 120 kHz in specific cases). This invalid setting shortens the OFDM symbol and slot duration, which directly reduces the physical-layer transmission time, HARQ round-trip time, and processing latency. Because the paper's headline metric is end-to-end delay, the reported values in Figs. 4 and 6 likely rely on an unsupported physical-layer configuration and are therefore not representative of any deployable 5G system in a factory. The paper should either use a valid FR1 numerology and rerun the simulations, or explicitly justify this configuration against the cited 3GPP specifications.","section":"§IV, Table I"},{"comment":"The physical environment is set to 'FlatGround' while the paper claims to evaluate the indoor factory profiles InF-SL, InF-DL, InF-SH, and InF-DH. FlatGround is not a 3GPP TR 38.901 InF environment, and with a single gNB and a single UE it does not capture the clutter, blockage, and shadowing statistics that distinguish the four InF profiles. The manuscript should explain how the FlatGround setting is combined with the InF path-loss models and demonstrate that the RF environment does not dominate the profile-specific differences shown in Fig. 3.","section":"§IV, Table I and Fig. 3"},{"comment":"No confidence intervals, number of simulation runs, random seed values, or simulation duration are reported. The paper's claims about latency stability and the differences between traffic classes and profiles are based on box plots, but the reader cannot determine whether those distributions come from enough independent replications to support the conclusions. This is load-bearing for the central claim and needs to be addressed, for example by reporting the number of replications and adding standard error or confidence-interval information.","section":"§IV, Figs. 4–6 and Table II"},{"comment":"The traffic classes NC, Video, and BE are described qualitatively, but the paper does not map them to concrete 3GPP or TSN latency, reliability, or cycle-time requirements for industrial control. Without such a mapping, the statement that 5G-TSN 'can effectively handle latency-critical applications in indoor factories' remains qualitative, and the abstract's claim that the findings 'demonstrate' the ability to address latency-sensitive scenarios is not quantitatively supported. The paper should state which latency budgets and reliability targets the simulated traffic is intended to meet.","section":"§III-B and §V"}],"minor_comments":[{"comment":"There is a typo in the text: 'the the highest SINR values' should be 'the highest SINR values'.","section":"§IV, Fig. 3"},{"comment":"Equations (1)–(8) are not numbered consistently in the text, and Eq. (6) is said to cover 'Eqs. 6, 7, 8' while the displayed equations do not have visible equation numbers. Please renumber and format the equations for clarity.","section":"§III-A"},{"comment":"The InF-HH profile is described in Section III-A but is not included in Table I or in the simulation results; please clarify whether it was excluded and why.","section":"§III-A and Table I"},{"comment":"Figure 6 shows only InF-SL and InF-SH, while the surrounding text discusses InF-DL and InF-DH in comparison; either add those profiles to the figure or adjust the text to match the presented data.","section":"§IV, Fig. 6"},{"comment":"The framework is listed as 'iNet 4.5.2' in the text; the official name is INET Framework. Please use the correct spelling.","section":"§IV"}],"recommendation":"major_revision","confidential_remarks":"The main concern is the invalid numerology setting, which affects the central latency results. The authors should be asked to rerun with a valid FR1 configuration. The paper also relies heavily on the authors' own prior survey [11] and earlier InF study [14] for motivation and novelty; the editor may wish to ask for an external comparison baseline before accepting the 'first to consider' claim."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Kouros and colleagues have put together a useful simulation study of 5G-TSN in an indoor factory setting. The genuinely new bit is the implementation of the 3GPP TR 38.901 indoor factory (InF) path-loss and LOS probability models inside Simu5G, and then using that to compare four InF profiles (SL, DL, SH, DH) for an AGV use case with three traffic classes. That is a reasonable and timely thing to do; the results showing that the dense-clutter profiles behave differently from sparse ones in terms of SINR, HARQ error rate, and end-to-end delay are plausible and instructive. The paper is clearly written and the authors are upfront that this is a controlled-environment study.\n\nThe soft spots are real, and one of them is load-bearing. Table I sets carrier frequency to 5.9 GHz (FR1) and numerology index to 4. In 3GPP NR, numerology index 4 means 240 kHz subcarrier spacing, which is not defined for FR1; FR1 supports 15, 30, and 60 kHz (indices 0-2). The stress-test note is correct: this invalid combination shortens the OFDM symbol duration and slot length, so the physical-layer transmission time, HARQ RTT, and processing delays are all compressed. Since the headline metric is end-to-end latency, the reported delays in Figs. 4 and 6 are likely unrealistically optimistic relative to any deployable FR1 system. This is not a matter of interpretation; the paper explicitly cites 3GPP TR 38.901/TS 38.211, and the configuration contradicts those specs. The authors need to rerun with a valid numerology (probably index 2 at 5.9 GHz, or justify a different band) and check whether the qualitative conclusions survive.\n\nThere are also secondary issues: no confidence intervals, seed counts, or run counts are reported for the box plots; there is no quantitative latency target to judge whether the values actually meet factory requirements (e.g., the IIC or 3GPP URLLC bounds); and no simulation artifacts are shared. The 'first to consider a wireless factory environment' claim is a bit strong given that the authors' own reference [14] already simulated InF environments, though that was channel characterization rather than full 5G-TSN.\n\nOn balance, this is a serious attempt at an important problem, and the InF implementation is a contribution in itself. The numerology error is fixable, but it affects the central quantitative claim, so the paper needs a major revision and a rerun before the results can be trusted. I would send it to peer review with the expectation of heavy revision, but I would not cite the current numerical results in my own work.","headline":"Useful 5G-TSN simulation study in 3GPP indoor factory channels, but an invalid FR1 numerology configuration undermines the headline latency numbers.","tokens_in":9877,"tokens_out":3032,"would_cite":false,"duration_ms":28512,"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 claims that a private 5G network integrated as a time-sensitive networking (TSN) bridge can carry a factory robot's high-priority control traffic with stable end-to-end latency across four standard indoor factory channel…","keywords":["5G","TSN","Wireless TSN","Industrial Networks","Indoor Factory","URLLC","End-to-end latency","HARQ"],"falsifier":"Repeat the same traffic mix in a real or more standard simulation setting with a common subcarrier spacing (for example 30 kHz at 3.5 GHz) and several UEs sharing the cell; if the high-priority control stream's end-to-end latency or jitter exceeds the bounds that TSN guarantees on wired Ethernet, the claim that 5G-TSN can address latency-sensitive factory scenarios would fail.","tokens_in":8827,"feed_emoji":"🏭","tokens_out":8838,"duration_ms":80302,"temperature":0.7,"pith_summary":"This paper sets out to show that wireless 5G can be fused with time-sensitive networking (TSN), the Ethernet mechanism that gives industrial traffic bounded latency, and still meet the timing demands of a factory floor. It simulates a private single-cell 5G network acting as a TSN bridge and carrying the traffic of a mobile industrial robot across four indoor factory channel profiles defined by the TR 38.901 report. The key result is that the highest-priority Network Control traffic keeps stable end-to-end latency across every profile and test case, while video and best-effort traffic show wider delay variation. The authors conclude that 5G-TSN can effectively handle latency-critical applications in indoor factories, at least in a controlled single-device setting.","feed_headline":"5G-TSN keeps factory control traffic stable in simulation","feed_subtitle":"A private 5G network acting as a TSN bridge holds steady latency for AGV control across four indoor factory profiles.","key_machinery":"The load-bearing object is the indoor factory channel model of the TR 38.901 report, implemented in the simulation with four profiles that vary clutter density and base-station height: sparse/low, dense/low, sparse/high, and dense/high. The model feeds radio propagation effects through path loss and line-of-sight probability into SINR, then into block-error probability and HARQ retransmissions, which is how the channel ultimately shapes end-to-end latency. On top of that radio core, the network classifies traffic into three TSN priority levels — Network Control, Video, and Best Effort — mapped to 5G quality-of-service bearers, so the simulation jointly measures the effects of radio environment, distance, and traffic priority.","core_discovery":"The central discovery is that, in the simulated 5G-TSN network, periodic Network Control traffic — packets of 50 to 500 bytes arriving every 50 ms to 1 s — keeps a stable end-to-end latency under all four indoor factory profiles (InF-SL, InF-DL, InF-SH, InF-DH), whereas Video and Best Effort flows show much wider delay spread. The simulations also show that InF-DL produces the lowest downlink SINR and InF-SH the highest, and that Hybrid Automatic Repeat reQuest (HARQ) error rates grow with distance in every profile, with InF-SL and InF-SH degrading most sharply and InF-DL and InF-DH tolerating longer links. The paper reads these results as evidence that 5G-TSN can address latency-sensitive scenarios in indoor factory environments.","pith_inferences":["I infer that the reported stability of control traffic reflects both TSN prioritization and the light load of 50- to 500-byte packets; the paper tests only one UE, so real multi-robot factories should be expected to see more HARQ retransmissions.","The configuration pairs numerology index 4 with a 5.9 GHz carrier, which is not a standard New Radio deployment; switching to a common subcarrier spacing could shift the absolute latency numbers even if the ranking across profiles remains.","The pattern that InF-SH has high SINR but worse long-range latency suggests high base stations are best deployed as small cells rather than as wide-coverage cells, an extension the paper does not draw itself.","If the results hold in multi-UE tests, they imply that safety-critical control loops can move onto the wireless 5G-TSN segment, leaving wired TSN for only the most stringent interlocks."],"forward_implications":["A single-cell private 5G network configured as a TSN bridge can keep periodic high-priority control traffic stable while video and best-effort flows share the same radio link.","Factory radio planning should treat the dense-clutter, low-base-station profile as the hardest case and the sparse-clutter, high-base-station profile as the best for short-range, high-rate links.","Distance is a first-order reliability factor: HARQ error rates rise in every profile as the terminal moves from 85 m to 255 m, so longer links should be assigned to the more latency-tolerant profiles.","The work provides a baseline for later 5G-TSN factory studies, since it is the first to combine the standard indoor factory profiles with a 5G-TSN network model."],"supporting_citations":[{"why":"Supplies the indoor factory channel profiles and path-loss and LOS-probability equations that define the four simulated environments.","marker":"[13]"},{"why":"Provides a simulation-and-validation study of 5G channels in indoor factory environments, grounding the realism of the channel model.","marker":"[14]"},{"why":"Provides the 5G New Radio simulator used to generate the SINR, latency, and HARQ results.","marker":"[20]"},{"why":"Provides the QoS-aware 5G-TSN simulation framework that maps TSN traffic classes onto 5G bearers.","marker":"[21]"},{"why":"Supplies the 5G-as-TSN-bridge configuration model that the simulated network architecture follows.","marker":"[12]"},{"why":"Prior integrated 5G-TSN performance study that motivates the feasibility of combining the two systems.","marker":"[2]"}],"fun_headline_variants":["5G-TSN keeps control latency stable in simulated factories","Network control traffic stays steady with 5G-TSN in four indoor profiles","Simulated 5G-TSN delivers stable latency for periodic control packets","5G-TSN holds control latency low across indoor factory scenarios"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The entire verdict rests on the simulation's radio settings matching real factory deployments: the configuration uses a single UE and single base station on flat ground, with numerology index 4 paired to a 5.9 GHz carrier, and if those choices are not representative of standard New Radio installations, the measured delays do not generalize.","fun_headline_variants_meta":{"raw":{"variants":["5G-TSN keeps control latency stable in simulated factories","Network control traffic stays steady with 5G-TSN in four indoor profiles","Simulated 5G-TSN delivers stable latency for periodic control packets","5G-TSN holds control latency low across indoor factory scenarios"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000156,"raw_usage":{"total_tokens":1199,"prompt_tokens":909,"completion_tokens":290,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":525,"completion_tokens_details":{"reasoning_tokens":215}},"tokens_in":525,"tokens_out":290,"duration_ms":3588,"temperature":1.0,"reasoning_tokens":215,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-10T16:47:24.104311+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Repeat the same traffic mix in a real or more standard simulation setting with a common subcarrier spacing (for example 30 kHz at 3.5 GHz) and several UEs sharing the cell; if the high-priority control stream's end-to-end latency or jitter exceeds the bounds that TSN guarantees on wired Ethernet, the claim that 5G-TSN can address latency-sensitive factory scenarios would fail.","supporting_citations":[{"cited_title":"5gtq: Qos-aware 5g-tsn simulation framework,","cited_arxiv_id":null,"evidence_quote":"Provides the QoS-aware 5G-TSN simulation framework that maps TSN traffic classes onto 5G bearers."},{"cited_title":"5g; study on channel model for frequencies from 0.5 to 100 ghz (3gpp tr 38.901 version 18.0.0 release 18),","cited_arxiv_id":null,"evidence_quote":"Supplies the indoor factory channel profiles and path-loss and LOS-probability equations that define the four simulated environments."},{"cited_title":"5g wireless channel characterization in indoor factory environments: Simulation and validation,","cited_arxiv_id":null,"evidence_quote":"Provides a simulation-and-validation study of 5G channels in indoor factory environments, grounding the realism of the channel model."},{"cited_title":"Simu5g: Simulator for 5g new radio networks,","cited_arxiv_id":null,"evidence_quote":"Provides the 5G New Radio simulator used to generate the SINR, latency, and HARQ results."},{"cited_title":"Work in progress: A centralized configuration model for tsn-5g networks,","cited_arxiv_id":null,"evidence_quote":"Supplies the 5G-as-TSN-bridge configuration model that the simulated network architecture follows."},{"cited_title":"Performance of integrated 3gpp 5g and ieee tsn networks,","cited_arxiv_id":null,"evidence_quote":"Prior integrated 5G-TSN performance study that motivates the feasibility of combining the two systems."}],"review_version":1}