{"id":"4a95a50b-56f8-486c-ad95-664962540196","arxiv_id":"2501.08159","paper_version":1,"verdict":"CONDITIONAL","confidence":"HIGH","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":4,"one_line_summary":"Sensing handover between network nodes can maintain continuous tracking of moving objects in ISAC networks, as shown by a three-base-station simulation.","lead":"This paper proposes that when a moving object is about to leave a base station's sensing view, the network should hand over the sensing task to a neighbor station, just as calls are handed over in cell networks. The idea matters for future 6G networks, which will combine communication with radar-like sensing and need to track objects as they move through dense cities.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The headline gain is computed by an oracle that instantly selects the best sensing configuration; without modeling measurement latency and handover overhead, the central quantitative claim is an optimistic upper bound.","rationale":"The reader's weakest_assumption identifies exactly the load-bearing idealization: instantaneous, cost-free detection and switching among all candidate sensing configurations. I agree with that identification, and the paper's own language in Section III confirms it. The phrase 'triggered as soon as a new ... configuration with better received sensing SNR is detected' is not a protocol description; it is an oracle that reports the upper envelope of all configuration SNRs. Every quantitative result in Fig. 3 and Table II follows from that envelope, so the central claim that mobility management 'reduces sensing interruption and boosts communication and sensing efficiency' is supported only under a strong feasibility assumption that the simulation never tests. The paper does contain a substantive conceptual contribution: it lays out triggering criteria, soft handover options, and the distinction between Tx-side and Rx-side handover, and it provides code for reproduction. Those are real merits. The concern is not that the idea is circular in the sense of being vacuous; it is that the simulated gain is partially built into the selection rule, and the unmodeled costs are not minor details. Sensing SNR depends on which APs have LOS to the object, and LOS can change on a faster timescale than handshaking and beam reconfiguration. A one- or two-snapshot latency could plausibly erase the claimed 50 dB drop avoidance in Fig. 3. Therefore the paper's current evidence supports a conditional claim: if the network can obtain and act on all configuration SNRs without delay or overhead, then the presented gains follow. Since the paper is positioned as a proof of concept rather than a system evaluation, conditional acceptance remains the right verdict, with the requested revision being a clear statement that the case study computes an upper bound and, ideally, a sensitivity check against latency and sounding overhead.","tokens_in":9293,"tokens_out":3182,"duration_ms":36136,"concrete_test":"Re-run the Section III case study with a non-oracle handover rule: at snapshot n, the candidate set is configurations whose SNR was measured or sounded within the last L snapshots, switching consumes T snapshots and a fraction of time/frequency resources, and measurement pilots interfere with communication. Vary L in {1,3,5} and T in {0,1,2}. If the minimum trajectory SNR falls below 10 dB, or if Table II success probabilities drop by more than a few percentage points, the central claim is not supported without modeling handover cost. A useful diagnostic is to plot the achieved SNR minus the envelope-max SNR; this gap should be non-positive and reveals the cost of realistic decisions.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The most load-bearing assumption is in Section III: 'the sensing handover is triggered as soon as a new mono-static or bi-static sensing configuration with better received sensing SNR is detected,' and 'the Tx/Rx functionality of the APs is adapted to maximize the sensing SNR.' This turns the handover rule into an envelope maximizer over exactly known, per-snapshot SNR values. Nothing in the model pays for obtaining those values: candidate configurations must be measured or predicted, which requires pilots, channel/object-state estimation, coordination between APs, and time; switching itself has handshaking latency and reconfiguration cost. The case study's flagship result—sustained sensing SNR above 10 dB, versus 50 dB drops without handover—is therefore an upper bound on what a real protocol could deliver, not an achieved performance. The same idealization inflates Table II's success probabilities and the conclusion's claim that the method 'reduces signaling overhead/network interference,' since signaling overhead is never counted. The conceptual contribution is real, and the qualitative direction likely survives, but the quantitative evidence for 'boosts efficiency' depends on removing the oracle.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper develops the concept of mobility management for integrated sensing and communication (ISAC) networks, focusing on sensing handover. It proposes a qualitative framework for triggering and executing sensing handover, including switches between monostatic and bistatic sensing and between transmitting/receiving nodes. The paper then presents a simulation case study with three access points, one moving object, and two UEs. The results show that, when handover is enabled, the sensing SNR in an interference-free scenario remains above 10 dB for the illustrated trajectory, whereas no fixed configuration can do so. In a second scenario with communication QoS constraints, handover restores UE SINR above the 15 dB threshold for most snapshots. The paper concludes that mobility management reduces sensing interruption, reduces signaling overhead, and improves communication and sensing efficiency.","tokens_in":9495,"tokens_out":5111,"duration_ms":52011,"significance":"The paper addresses an important and timely problem: how to maintain continuous sensing in ISAC networks under mobility and blockage. Its conceptual contributions—classifying handover triggers, describing soft sensing handover, and switching sensing modalities—are useful for the emerging 6G/ISAC literature. The case study is clearly described, the simulation parameters are explicit, and the authors provide reproducible simulation code on GitHub. The communication-QoS-triggered handover in Fig. 4 is a legitimate proof of concept. However, the quantitative evidence for the headline efficiency claims is weakened by an oracle-style handover rule in Fig. 3 and by the absence of any overhead or latency model. The conceptual framing is strong; the empirical validation is currently an upper-bound idealization.","major_comments":[{"comment":"The handover rule in Section III states that 'the sensing handover is triggered as soon as a new mono-static or bi-static sensing configuration with better received sensing SNR is detected' and that 'the Tx/Rx functionality of the APs is adapted to maximize the sensing SNR.' Under this rule, the 'with handover' curve in Fig. 3 is the pointwise maximum (upper envelope) of the individual configuration curves. Consequently, the headline result—sustained sensing SNR above 10 dB and the avoidance of a 50 dB drop—is a direct consequence of the selection rule, not evidence of a protocol-level gain. To support the paper's central claim, the authors need to model how the network measures or predicts the achievable SNR of every candidate configuration (including pilots, coordination, and handover latency) and compare a realistic handover policy against a baseline that lacks this oracle information. As written, the case study establishes an upper bound rather than an achieved performance.","section":"Section III, Fig. 3"},{"comment":"The abstract promises sensing 'with low overhead,' and the conclusions state that sensing handover 'reduces signaling overhead/network interference,' but no overhead, latency, or signaling cost is modeled anywhere in the simulation. Each handover in Fig. 3 and Fig. 4 requires evaluating candidate configurations, exchanging assisting information between APs, and reconfiguring beams; none of these costs appears in the curves or in Table II. The 'boosts communication and sensing efficiency' claim is therefore not quantitatively supported. The authors should either add an overhead and latency model with explicit results, or restrict their claims to sensing continuity and interference avoidance for a zero-cost handover idealization.","section":"Abstract, Section III (Fig. 4, Table II), Section IV"},{"comment":"Table II reports success probabilities without confidence intervals, standard deviations, or the number of Monte Carlo iterations. The results are also surprisingly insensitive in the relaxed-threshold regime: at a 5 dB sensing SINR threshold, handover-enabled and handover-disabled success probabilities differ by only 0.1 percentage points, and the expected number of handovers is 0.002 and 0.004 for the 5 dB and 10 dB thresholds, respectively. This means that with random UE positions, the handover mechanism is rarely exercised except under strict sensing requirements. The single trajectory in Fig. 3/Fig. 4 and the unquantified random-UE averages in Table II make it difficult to assess how representative the claimed gains are. Please provide a distribution of outcomes or confidence intervals, and discuss the representativeness of the illustrated trajectories.","section":"Table II"}],"minor_comments":[{"comment":"The statement that sensing handover 'guarantees uninterrupted sensing' is stronger than the simulation evidence: Fig. 4 shows snapshots where UE1's SINR remains below threshold even after handover, and Table II reports success probabilities below 100%. Suggest rewording to 'enables near-continuous sensing in the considered scenarios.'","section":"Section IV"},{"comment":"The case study only models mobility of the sensed object; the APs and UEs are stationary (the UEs are explicitly described as stationary in Fig. 4). Since the introduction motivates mobile transceivers and UEs, please clarify that the numerical evaluation covers only the moving-object scenario and leave mobile-transceiver mobility for future work.","section":"Section III"},{"comment":"The sentence 'The handover SNR curve indicates the sensing configuration' is ambiguous; it likely means the handover curve is the achieved SNR rather than a configuration indicator. Please rephrase for clarity.","section":"Fig. 3 description"},{"comment":"The self-interference power is set to −45 dBm, which is 85 dB below the 40 dBm transmit power and only 15 dB above the −60 dBm noise power. The paper does not justify this SI cancellation level; a sensitivity analysis over SI cancellation quality would strengthen the results.","section":"Table I"},{"comment":"The list of triggering events is helpful, but the paper would benefit from a compact decision flowchart connecting the various triggers to the handover procedure, especially since the simulation implements only the 'better SNR' and 'QoS threshold' triggers from the full taxonomy.","section":"Section II"}],"recommendation":"major_revision","confidential_remarks":"The paper is a concept/position paper with an illustrative case study, not a complete protocol design. Its main strengths are the clear conceptual framing and the reproducible simulation code. The principal risk is the mismatch between the oracle-style handover rule in the simulations and the claims of low overhead and improved efficiency. If the authors can either add a realistic overhead model or carefully reframe the results as an ideal-case upper bound, the paper would be acceptable for publication after revision."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Two things to know. First, this is the first paper I've seen that treats sensing handover as a distinct mobility-management problem in ISAC. Second, its headline result—sustained sensing SNR above 10 dB instead of 50 dB drops—comes from a handover rule that assumes the network knows every candidate configuration's SNR at every snapshot and can switch for free. That makes the quantitative claim an optimistic upper bound, not an achieved protocol performance.\n\nWhat's genuinely new: the trigger taxonomy (coverage, beam restrictions, resource allocation, blockage, interference, resolution) is well organized and covers the right cases. The monostatic/bistatic switching and the Tx/Rx handover distinction are clearly explained. Soft handover for sensing is a sensible extension of the communication concept. The simulations are simple but reproducible—they actually ship the code on GitHub, which is credit-worthy. Figure 3 demonstrates that, under the oracle assumption, handover keeps the SNR above 10 dB, and Figure 4 shows the idea can protect UE QoS by moving the sensing link away from a UE. As a proof-of-concept, the idea is coherent and the direction is plausible.\n\nThe soft spots are where the reader and the stress-test note point. The 'as soon as a better configuration is detected' rule is a free lunch: measuring all candidate configurations requires pilots, estimation, and coordination; switching requires handshaking and reconfiguration. None of that is modeled, so the claimed gains are an envelope, not a result. The abstract and conclusions go further and claim the method reduces signaling overhead, which is unsupported—the opposite is more likely in a real system. Table II lacks confidence intervals or iteration counts, so the 99.9% numbers aren't statistically grounded. The baseline for 'handover disabled' is mono-static with one AP, which is weak; a stronger baseline would be the best fixed configuration or a communication-only handover. The circularity concern is real but not fatal: of course the handover curve is the max of the candidates; the open question is what it costs to know which candidate is best.\n\nNet: the paper deserves a serious referee. The concept is new, the writing is clear, and the code is a plus. Revisions should either model the measurement and switching overhead or explicitly frame the results as an upper bound. I'd bring it to a reading group for the concept discussion, and I'd cite it as the first explicit sensing-handover framework.","headline":"First explicit sensing-handover framework for ISAC; useful concept and clean writing, but the headline SNR gain is an oracle envelope until measurement and switching costs are modeled.","tokens_in":10029,"tokens_out":2539,"would_cite":true,"duration_ms":25530,"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":"Sensing handover keeps object tracking above 10 dB SNR","keywords":["mobility management","integrated sensing and communications","sensing handover","monostatic sensing","bistatic sensing","interference management","beamforming","6G"],"falsifier":"Measure, in a small testbed of three cooperating access points tracking a moving object, the real time from a handover trigger to a fully reconfigured bistatic or monostatic sensing link; if that latency spans even a few snapshots while the object is in the coverage gap, the sensing SNR will drop below $10$ dB and the uninterrupted-sensing claim fails. Equivalently, a field test with random user positions that includes handover signaling overhead could show the $99.2\\%$ success rate for the $15$ dB threshold dropping once measurement latency is counted.","tokens_in":9087,"feed_emoji":"📡","tokens_out":11160,"duration_ms":102911,"temperature":0.7,"pith_summary":"Integrated sensing and communication (ISAC) networks use the same wireless infrastructure to communicate and to track objects, and both jobs are sensitive to movement: a tracked object or a transceiver can lose the line-of-sight link that radar-like sensing depends on, and moving targets can shift interference onto communication links. This paper argues that the network should treat sensing as a handoverable service, transferring the sensing transmitter and/or receiver role to neighboring access points and switching between monostatic sensing (one node transmits and receives) and bistatic sensing (two cooperating nodes) as conditions change. The central case study tracks one object with three cooperating access points and shows that sensing handover keeps the sensing SNR above $10$ dB for the whole trajectory, while no single fixed configuration can sense the object without interruption. With communicating user devices added, handover also restores user SINR when the sensed object aligns with a communication link, keeping quality-of-service while improving sensing. If this holds, mobility management is a practical, low-overhead way to make ISAC sensing continuous in the dense multi-node networks expected in 6G.","feed_headline":"Sensing handover keeps object tracking above 10 dB","feed_subtitle":"In networks that combine radar and communication, switching the sensing role between nodes avoids dropouts.","key_machinery":"The load-bearing mechanism is the sensing handover procedure, defined as the transfer of the sensing transmitter and/or receiver role from a source access point to target access point(s), with the sensing modality chosen anew between monostatic (one node transmits and receives) and bistatic (two cooperating nodes) configurations. The procedure has two stages: determining the need for handover from predefined criteria—long-term conditions such as limited sensing coverage area and beam restrictions, plus dynamic events such as resource-allocation trade-offs, blockage, interference management, and sensing resolution—and performing the handover by sending assisting information (object trajectory, position, available resources, beam restrictions, coverage maps) to the target node before it accepts and reconfigures. A soft handover option keeps the source node sensing for an overlap period. In the case study, the deciding quantity is the received sensing SNR of every monostatic and bistatic configuration at each snapshot; the network switches to the configuration with the best SNR, and when communication users are present it also switches when the user SINR falls below the QoS threshold. This mechanism converts mobility and interference events into configuration changes, which is what allows continuous sensing.","core_discovery":"The paper's central claim is that sensing handover—moving the transmission and/or reception of sensing signals to neighbor nodes and updating whether sensing is monostatic or bistatic—eliminates the sensing interruptions caused by mobility in ISAC networks. In the paper's simulation of three access points tracking one moving object, the handover is triggered as soon as a candidate monostatic or bistatic configuration offers a better received sensing SNR. This event-driven switching produces a sensing SNR that stays above $10$ dB across the simulated trajectory, with handovers at four points; without handover, none of the individual configurations can provide uninterrupted sensing, and one switch to a new node avoids a $50$ dB SNR drop. When two communicating users are added, handover is triggered when the user SINR falls below the $15$ dB requirement, and the results show both users' requirements are met for most of the trajectory while sensing performance improves. Averaged over random user positions, handover keeps sensing SINR above thresholds of $5$, $10$, and $15$ dB with probabilities $99.9\\%$, $99.9\\%$, and $99.2\\%$, versus $99.8\\%$, $91.3\\%$, and $60.4\\%$ without handover, using an average of at most about one handover per trajectory.","pith_inferences":["A practical network must estimate each candidate configuration's sensing SNR before deciding, and the paper assumes this information is available at every snapshot with instantaneous, cost-free switching; measurement latency and signaling overhead are therefore not included in the reported gains.","Adding trajectory prediction to the handover trigger would make the switch proactive rather than reactive and would likely remove the brief SNR dips seen at handover points; this is a directly testable modification of the case study.","The same event-based logic could be applied per object in multi-object tracking, at the cost of coordinating beams and resources across simultaneous sensing handovers, which the paper does not analyze."],"forward_implications":["A moving object can be tracked continuously along its trajectory as long as some cooperating node or node pair retains line of sight, even when no single configuration does for the whole path.","Interference events can be used as handover triggers, so communication quality-of-service (user SINR above $15$ dB in the case study) can be maintained while sensing continues.","In dense multi-node deployments, the cost is modest: for the simulated random user positions, the average number of handovers per trajectory stays near or below one, even for a strict $15$ dB sensing SINR requirement.","Handover between monostatic and bistatic configurations bridges coverage gaps between nodes, including a case where switching avoids a $50$ dB sensing SNR drop."],"supporting_citations":[{"why":"Frames ISAC as perceptive mobile networks where nodes, user devices, and sensed objects are mobile, motivating mobility management as an ISAC design problem.","marker":"[9]"},{"why":"Provides measured line-of-sight probabilities that sensing depends on, the core reason handover is needed when links are lost intermittently.","marker":"[11]"},{"why":"The cellular physical-layer specification whose procedures the sensing handover reconfiguration adapts from communication handover.","marker":"[13]"},{"why":"The cellular radio-resource-control specification that defines communication handover signaling, the baseline the proposed sensing handover modifies.","marker":"[14]"},{"why":"Public simulation code for reproducing the case-study results that carry the paper's central claim.","marker":"[15]"},{"why":"The standardization study item on ISAC, cited to show sensing handover procedures do not yet exist in standards.","marker":"[8]"}],"fun_headline_variants":["Handover in ISAC networks prevents sensing dropouts","Sensing handover sustains 10 dB SNR for moving objects","Mobility management boosts ISAC sensing performance","Switching sensing tasks between nodes avoids interruptions","Event-driven handover keeps ISAC tracking above 10 dB"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The paper's simulated gains assume the network knows the sensing SNR of every candidate monostatic and bistatic configuration at each snapshot and can switch to the best one instantly at no cost.","fun_headline_variants_meta":{"raw":{"variants":["Handover in ISAC networks prevents sensing dropouts","Sensing handover sustains 10 dB SNR for moving objects","Mobility management boosts ISAC sensing performance","Switching sensing tasks between nodes avoids interruptions","Event-driven handover keeps ISAC tracking above 10 dB"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000432,"raw_usage":{"total_tokens":2248,"prompt_tokens":1031,"completion_tokens":1217,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":647,"completion_tokens_details":{"reasoning_tokens":1140}},"tokens_in":647,"tokens_out":1217,"duration_ms":9807,"temperature":1.0,"reasoning_tokens":1140,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-10T20:28:45.847733+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Measure, in a small testbed of three cooperating access points tracking a moving object, the real time from a handover trigger to a fully reconfigured bistatic or monostatic sensing link; if that latency spans even a few snapshots while the object is in the coverage gap, the sensing SNR will drop below $10$ dB and the uninterrupted-sensing claim fails. Equivalently, a field test with random user positions that includes handover signaling overhead could show the $99.2\\%$ success rate for the $15$ dB threshold dropping once measurement latency is counted.","supporting_citations":[{"cited_title":"Perceptive mobile networks: Cellular networks with radio vision via joint communication and radar sensing,","cited_arxiv_id":null,"evidence_quote":"Frames ISAC as perceptive mobile networks where nodes, user devices, and sensed objects are mobile, motivating mobility management as an ISAC design problem."},{"cited_title":"Multistatic sensing performance maps for evaluating integrated sensing and communication deployments,","cited_arxiv_id":null,"evidence_quote":"Provides measured line-of-sight probabilities that sensing depends on, the core reason handover is needed when links are lost intermittently."},{"cited_title":"(2024, Jul) Technical specification 38.214","cited_arxiv_id":null,"evidence_quote":"The cellular physical-layer specification whose procedures the sensing handover reconfiguration adapts from communication handover."},{"cited_title":"(2024, Jul) Technical specification 38.331","cited_arxiv_id":null,"evidence_quote":"The cellular radio-resource-control specification that defines communication handover signaling, the baseline the proposed sensing handover modifies."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Public simulation code for reproducing the case-study results that carry the paper's central claim."},{"cited_title":"(2023) Study on integrated sensing and communication","cited_arxiv_id":null,"evidence_quote":"The standardization study item on ISAC, cited to show sensing handover procedures do not yet exist in standards."}],"review_version":1}