{"id":"6b157949-e1d1-4d2e-bd09-af699b17bef1","arxiv_id":"2605.14920","paper_version":1,"verdict":"UNVERDICTED","confidence":"LOW","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"FU-MPC is a receding-horizon MPC controller that treats motorized LiDAR rotation as an explicit variable to jointly maximize frontier exploration utility and minimize direction-dependent localization uncertainty during UAV flight.","lead":"The paper describes a UAV system with a motorized rotating LiDAR and a hierarchical planner using FU-MPC to optimize sensor rotation for both faster frontier coverage and lower localization uncertainty. A smart generalist might read it for insight into how adding an independent sensing degree of freedom can improve autonomous mapping without extra drone maneuvers.","discovery_kind":"new_method","skeptic_critique":{"model":"grok-4.3","headline":"Lightweight surrogate approximation of joint objective lacks explicit validation against full evaluation","rationale":"The reader's weakest assumption matches the load-bearing point exactly. The abstract-only review already flags the surrogate's fidelity as critical; absent full-text details showing validation (e.g., via ablation or correlation), this remains the least secure link for the experimental demonstration. No other internal inconsistency appears from the given material.","tokens_in":1746,"tokens_out":277,"duration_ms":13244,"concrete_test":"On the experimental trajectories, recompute the joint objective using the full (non-surrogate) evaluation; if the surrogate-selected actions differ in >15% of cases or the efficiency/localization deltas versus baselines shrink below statistical significance, the approximation undermines the claim.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim rests on FU-MPC delivering improved exploration efficiency and robust localization via real-time optimization of LiDAR rotation. This depends on the surrogate evaluation (mentioned in abstract) sufficiently capturing the combined frontier utility and direction-dependent uncertainty without significant degradation. If the surrogate introduces approximation error that alters selected trajectories relative to the true objective, the reported gains over fixed-pattern and uncertainty-only baselines may not hold in the tested complex environments. No independent check (e.g., correlation or full-objective re-ranking) is referenced in the provided abstract.","agreement_with_reader":"agree"},"referee_report":{"model":"grok-4.3","summary":"The paper proposes a hierarchical UAV exploration system with a motorized rotating LiDAR. A global planner sequences topology-aware frontier viewpoints, while the local FU-MPC controller optimizes LiDAR rotation over a receding-horizon trajectory by jointly maximizing frontier-aware exploration utility and minimizing direction-dependent localization uncertainty. A lightweight surrogate enables real-time onboard execution of this optimization. Experiments in complex environments report improved exploration efficiency and maintained localization robustness relative to fixed-pattern scanning and uncertainty-only baselines.","tokens_in":1850,"tokens_out":314,"duration_ms":19857,"significance":"If the experimental gains are reproducible and the surrogate approximation is shown to preserve trajectory quality, the work would provide a concrete demonstration that decoupling sensor orientation from UAV motion can improve the exploration-localization trade-off in geometrically challenging settings. The integration of frontier utility with uncertainty in an MPC formulation for active LiDAR control is a practical step toward more adaptive sensing platforms.","major_comments":[{"comment":"Abstract (final paragraph) and the description of the surrogate evaluation: the central claim that FU-MPC improves efficiency while preserving localization rests on the lightweight surrogate accurately capturing the joint objective without materially altering the optimized trajectories relative to a full evaluation. No correlation analysis, full-objective re-ranking of candidate trajectories, or ablation quantifying approximation error is referenced, which directly affects whether the reported gains over the two baselines can be attributed to the proposed joint objective rather than to the surrogate itself.","section":"Abstract"}],"minor_comments":[],"recommendation":"major_revision","confidential_remarks":null},"author_rebuttal":{"model":"grok-4.3","summary":"We thank the referee for the constructive feedback. We address the major comment below and will revise the manuscript to strengthen the validation of the surrogate approximation.","responses":[{"response":"We agree that explicit validation of the surrogate is needed to support attribution of the reported gains to the joint objective. In the revised version we will add (i) a correlation analysis between surrogate and full-objective scores over a set of candidate trajectories, (ii) a re-ranking experiment showing how often the surrogate selects the same top trajectory as the full objective, and (iii) an ablation quantifying the approximation error and its effect on final exploration and localization metrics. These results will be placed in the surrogate-model section and referenced from the abstract and experimental discussion.","revision_made":"yes","referee_comment":"[Abstract] Abstract (final paragraph) and the description of the surrogate evaluation: the central claim that FU-MPC improves efficiency while preserving localization rests on the lightweight surrogate accurately capturing the joint objective without materially altering the optimized trajectories relative to a full evaluation. No correlation analysis, full-objective re-ranking of candidate trajectories, or ablation quantifying approximation error is referenced, which directly affects whether the reported gains over the two baselines can be attributed to the proposed joint objective rather than to the surrogate itself."}],"tokens_in":1367,"tokens_out":279,"duration_ms":12060,"standing_objections":[]},"desk_editor":{"model":"grok-4.3","letter":"The paper's contribution is a receding-horizon controller (FU-MPC) that adds LiDAR rotation angle as a decision variable inside the optimization, linking it directly to both frontier utility and localization uncertainty. The global layer sequences viewpoints with topology costs, then the local MPC plans the scan pattern along the predicted trajectory. A lightweight surrogate is used so the whole thing runs onboard.\n\nThat decoupling of sensor direction from platform motion is a practical step for UAVs that already carry actuated LiDARs. The hierarchical split and the joint objective are reasonable engineering choices, and the claim that it improves efficiency over fixed-pattern and uncertainty-only baselines is at least plausible.\n\nThe main weakness is the surrogate. The abstract states it enables real-time execution but supplies no correlation check, re-ranking experiment, or comparison against the full objective, so it is unclear whether the optimized rotations are close to what the true utility would select. The experimental claims also lack visible error bars, run counts, or statistical tests, which makes it hard to judge how consistent the reported improvements actually are.\n\nThis is aimed at groups already working on active-sensing UAV exploration who can replicate the hardware. It is not foundational, but the formulation is specific enough that a referee could check whether the surrogate approximation holds and whether the baselines are fairly implemented. I would send it to review.","headline":"The new element is treating motorized LiDAR rotation as an explicit MPC variable that trades off frontier coverage against direction-dependent uncertainty, but the abstract gives no evidence the surrogate matches the true objective or that the gains are statistically reliable.","tokens_in":2348,"tokens_out":359,"would_cite":false,"duration_ms":15399,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"grok-4.3","headline":"Optimizing motorized LiDAR rotation along UAV paths increases exploration speed while preserving localization accuracy.","keywords":["UAV exploration","motorized LiDAR","model predictive control","frontier detection","localization uncertainty","receding horizon control"],"falsifier":"Deploy the system in a geometrically challenging indoor or outdoor scene and measure whether explored volume per unit time fails to exceed the fixed-pattern baseline or if pose estimation error rises above the uncertainty-only baseline.","tokens_in":2663,"feed_emoji":"🚁","tokens_out":564,"duration_ms":27379,"temperature":0.7,"pith_summary":"The paper presents a UAV equipped with an independently actuated rotating LiDAR and a hierarchical framework to explore unknown spaces. A global planner sequences representative viewpoints using topology-aware costs. The local FU-MPC controller then optimizes the LiDAR rotation angle over a receding horizon by balancing frontier coverage utility against direction-dependent localization uncertainty. Lightweight surrogate evaluation keeps the optimization fast enough for real-time onboard execution. Experiments in complex environments show higher efficiency than fixed-pattern scanning or uncertainty-only baselines without loss of localization robustness.","feed_headline":"Motorized LiDAR rotation speeds UAV exploration","feed_subtitle":"Controller optimizes sensor angle to cover frontiers faster while controlling uncertainty growth.","key_machinery":"Frontier- and Uncertainty-Aware Model Predictive Control (FU-MPC), which optimizes the motorized LiDAR rotation angle as a decision variable in the receding-horizon objective that combines frontier utility and localization uncertainty costs.","core_discovery":"Treating LiDAR rotation as an explicit decision variable in model predictive control enables joint optimization of exploration progress and localization quality by maximizing frontier-aware utility while penalizing rotations that increase direction-dependent uncertainty, with surrogate models supporting real-time execution on the actuated sensor platform.","pith_inferences":["The decoupling of sensing direction from vehicle motion could extend to other mobile platforms with actuated sensors.","Integration with global planners that also penalize uncertainty might further reduce drift in long missions.","Scaling the surrogate accuracy with environment size could become a bottleneck in very large spaces."],"forward_implications":["Coverage expands with fewer UAV translational or rotational maneuvers because sensor direction is adjusted independently.","Localization stays reliable because the controller explicitly accounts for how scan direction affects uncertainty.","Real-time onboard operation remains feasible through surrogate evaluation of the combined objective.","Performance exceeds both fixed-pattern scanning and uncertainty-only approaches in complex environments."],"fun_headline_variants":["FU-MPC optimizes LiDAR rotation for UAV exploration progress","Frontier-aware MPC with uncertainty terms for rotating LiDAR","Hierarchical exploration using MPC to control motorized LiDAR angle","UAV MPC controller factors LiDAR direction into localization quality"],"cache_read_input_tokens":2112,"weakest_assumption_plain":"The lightweight surrogate evaluation accurately captures the joint exploration-uncertainty objective for the full horizon without causing the optimized trajectories to degrade coverage or localization quality.","fun_headline_variants_meta":{"raw":{"variants":["FU-MPC optimizes LiDAR rotation for UAV exploration progress","Frontier-aware MPC with uncertainty terms for rotating LiDAR","Hierarchical exploration using MPC to control motorized LiDAR angle","UAV MPC controller factors LiDAR direction into localization quality"]},"model":"grok-4.3","cost_usd":0.006801,"raw_usage":{"total_tokens":3167,"prompt_tokens":678,"num_sources_used":0,"completion_tokens":65,"cost_in_usd_ticks":68012000,"prompt_tokens_details":{"text_tokens":678,"audio_tokens":0,"image_tokens":0,"cached_tokens":256},"completion_tokens_details":{"audio_tokens":0,"reasoning_tokens":2424,"accepted_prediction_tokens":0,"rejected_prediction_tokens":0}},"tokens_in":678,"tokens_out":65,"duration_ms":22056,"temperature":1.0,"reasoning_tokens":2424,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-06-30T20:36:25.197777+00:00","model_set":{"reader":"grok-4.3"},"falsifier":"Deploy the system in a geometrically challenging indoor or outdoor scene and measure whether explored volume per unit time fails to exceed the fixed-pattern baseline or if pose estimation error rises above the uncertainty-only baseline.","supporting_citations":[],"review_version":1}