{"id":"22b2d152-a9d3-4b3f-8e3e-297206b2b5e5","arxiv_id":"2605.15074","paper_version":1,"verdict":"UNVERDICTED","confidence":"LOW","novelty_score":6.0,"correctness_risk":"unknown","formal_verification":"none","parameter_count":0,"one_line_summary":"SOCC-ICP unifies semantic occupancy grid mapping with LiDAR scan alignment via adaptive point-to-point or point-to-plane ICP, achieving competitive odometry performance while producing maps usable for planning.","lead":"The paper presents SOCC-ICP, a framework that performs LiDAR odometry and semantic occupancy grid mapping together using adaptive ICP on voxels that store both geometry and semantics. This single map structure is intended to support both pose estimation and downstream tasks like motion planning without maintaining separate representations.","discovery_kind":"new_method","skeptic_critique":{"model":"grok-4.3","headline":"The unification claim rests on raycasting reliably filtering dynamics and local planarity statistics correctly selecting ICP modes; both are unverified in mixed or degenerate scenes.","rationale":"The reader's weakest_assumption exactly identifies the two internal assumptions required for the central unification claim; the abstract-only limitation noted by the reader is the reason the verdict remains UNVERDICTED rather than a stronger stance.","tokens_in":1698,"tokens_out":342,"duration_ms":18242,"concrete_test":"On a sequence containing both static and moving objects (e.g., KITTI 07 with synthetic dynamic overlays), disable semantic cues and compare ATE when raycasting is enabled vs. disabled; also ablate the planarity-based ICP selector against fixed point-to-plane. If ATE rises >15% with raycasting or the selector, the filtering/selection assumptions do not hold.","verdict_should_be":"UNCHANGED","load_bearing_attack":"For the single-representation claim to hold, the occupancy grid must simultaneously support accurate ICP-based odometry and produce a map usable downstream. The method encodes per-voxel geometric/semantic stats and uses raycasting for free-space updates to remove dynamics, then selects point-to-point vs. point-to-plane ICP via local planarity. If raycasting marks dynamic points as free space incorrectly (e.g., under partial occlusion or fast motion) or if planarity thresholds misclassify mixed surfaces, alignment errors propagate into both the pose estimate and the map itself. The abstract asserts robustness even without semantics, yet provides no quantitative isolation of these two mechanisms, leaving the dual-use premise dependent on assumptions whose failure would simultaneously degrade odometry and map quality.","agreement_with_reader":"agree"},"referee_report":{"model":"grok-4.3","summary":"The manuscript introduces SOCC-ICP, a semantics-assisted LiDAR odometry framework that performs joint semantic occupancy grid mapping and scan alignment within a single voxel-based representation. Each voxel encodes geometric and semantic statistics to support adaptive selection between point-to-point and point-to-plane ICP according to local planarity, while raycasting-based free-space updates are used to filter dynamic objects. The paper claims that this unified approach achieves performance competitive with state-of-the-art LiDAR odometry methods across diverse scenarios, remains robust in geometrically degenerate environments even without semantic cues, yields further gains when semantics are available, and directly supplies a map usable for downstream tasks without redundant structures.","tokens_in":1850,"tokens_out":580,"duration_ms":28287,"significance":"If the empirical claims hold, the work would be significant for robotics by addressing the typical separation between odometry representations (points, surfels, NDT) and planning representations (occupancy grids). A single structure that supports both accurate pose estimation and downstream usability could reduce system complexity. The adaptive ICP mechanism and dynamic filtering via raycasting are conceptually appealing, but their contribution to the unification claim requires concrete validation that is not evident from the provided description.","major_comments":[{"comment":"Abstract and §5 (Evaluation): the central claim of 'performance competitive with state-of-the-art' and 'robustness in geometrically degenerate environments' is asserted without any reported error metrics (RMSE, ATE, RPE), dataset names, sequence counts, or quantitative baseline comparisons. This absence prevents assessment of whether the unification actually delivers the stated benefits.","section":"Abstract and §5"},{"comment":"§3 (Method): the unification premise rests on two unverified mechanisms—raycasting reliably marking dynamic points as free space and local planarity statistics correctly selecting ICP modes. No ablation studies, failure-case analysis under partial occlusion or fast motion, or sensitivity tests on planarity thresholds are described; failure of either mechanism would simultaneously degrade both odometry and map quality.","section":"§3"}],"minor_comments":[{"comment":"The description of voxel encoding for geometric/semantic statistics lacks explicit notation or pseudocode, making it difficult to reproduce the per-voxel update rules.","section":"§3.1"},{"comment":"No mention of computational overhead (memory per voxel, raycasting cost) relative to separate odometry + mapping pipelines, which would be needed to substantiate the 'eliminates redundant map structures' claim.","section":"§4"}],"recommendation":"major_revision","confidential_remarks":"The manuscript appears to be a methods contribution whose central claims cannot be evaluated from the supplied text; the editor may wish to request the full experimental tables and code before further review."},"author_rebuttal":{"model":"grok-4.3","summary":"We thank the referee for the constructive feedback. The comments correctly identify areas where the manuscript requires additional quantitative detail and validation to support its claims. We address each point below and will revise accordingly.","responses":[{"response":"We agree that explicit quantitative results are necessary to substantiate the performance claims. The revised manuscript will include RMSE, ATE, and RPE metrics on standard datasets (e.g., KITTI sequences with specified counts), along with direct numerical comparisons to baselines such as LOAM and LeGO-LOAM. This will enable assessment of the unification benefits.","revision_made":"yes","referee_comment":"[Abstract and §5] Abstract and §5 (Evaluation): the central claim of 'performance competitive with state-of-the-art' and 'robustness in geometrically degenerate environments' is asserted without any reported error metrics (RMSE, ATE, RPE), dataset names, sequence counts, or quantitative baseline comparisons. This absence prevents assessment of whether the unification actually delivers the stated benefits."},{"response":"The referee is correct that the mechanisms require explicit validation. We will add ablation studies quantifying the impact of raycasting-based free-space updates for dynamic filtering and the planarity-driven ICP mode selection. The revision will also include failure-case analysis under occlusion/fast motion and sensitivity tests on the planarity threshold.","revision_made":"yes","referee_comment":"[§3] §3 (Method): the unification premise rests on two unverified mechanisms—raycasting reliably marking dynamic points as free space and local planarity statistics correctly selecting ICP modes. No ablation studies, failure-case analysis under partial occlusion or fast motion, or sensitivity tests on planarity thresholds are described; failure of either mechanism would simultaneously degrade both odometry and map quality."}],"tokens_in":1412,"tokens_out":384,"duration_ms":20061,"standing_objections":[]},"desk_editor":{"model":"grok-4.3","letter":"The core idea is to store geometric and semantic stats per voxel, pick point-to-point or point-to-plane ICP from local planarity, and use raycasting to mark free space and drop dynamics. That single map then feeds downstream tasks without a second representation.\n\nThe approach is straightforward and directly targets the usual split between odometry maps and occupancy grids. Encoding both geometry and semantics in the same voxels is a clean way to let semantics influence downsampling and correspondence weighting when labels are present.\n\nThe soft spot is the complete absence of results. No error metrics, no datasets, no comparison tables, and no ablation on the planarity test or the raycasting step appear in the text. The claim of competitive performance and robustness in degenerate scenes therefore rests on unshown evidence. The stress-test concern is fair: if raycasting mislabels dynamics under occlusion or the planarity threshold picks the wrong ICP mode on mixed surfaces, both the pose and the map degrade together, yet nothing demonstrates these mechanisms hold.\n\nThe paper is aimed at mobile-robotics teams that already maintain occupancy grids and want to collapse the localization layer into them. Someone implementing LiDAR pipelines might borrow the voxel statistic layout even if they keep a separate odometry map.\n\nIf the full manuscript contains standard-dataset experiments with clear baselines and ablations, it is worth sending to review. Without those numbers the contribution stays at the level of a framework sketch.","headline":"SOCC-ICP tries to run ICP odometry inside a semantic occupancy grid so the same structure serves both localization and planning, but the abstract supplies no numbers to check whether the unification actually works.","tokens_in":2328,"tokens_out":371,"would_cite":false,"duration_ms":20546,"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":"SOCC-ICP performs LiDAR odometry and semantic occupancy grid mapping inside one shared voxel structure.","keywords":["LiDAR odometry","occupancy grid","ICP","semantic mapping","dynamic object filtering","robot navigation"],"falsifier":"Odometry error rising above baseline methods on a sequence containing many independently moving objects that raycasting fails to clear from the grid.","tokens_in":2597,"feed_emoji":"🗺️","tokens_out":603,"duration_ms":17029,"temperature":0.7,"pith_summary":"The paper introduces a method that builds a semantic occupancy grid during LiDAR scan alignment rather than maintaining separate structures for each task. Each voxel stores both geometric statistics and semantic labels, which lets the system switch between point-to-point and point-to-plane ICP according to local planarity and removes dynamic objects through raycasting free-space updates. The same grid is then available for motion planning and other downstream uses without extra conversion steps. When semantic labels are present they improve correspondence weighting and downsampling, but the system still works competitively without them. A reader would care because the approach removes the usual duplication of map data between odometry and planning modules.","feed_headline":"One voxel grid runs both LiDAR odometry and semantic mapping","feed_subtitle":"SOCC-ICP stores geometry and semantics per cell to adapt ICP and clear dynamic objects, yielding a planning-ready map without extra structur","key_machinery":"The semantic occupancy grid voxel that stores both geometric and semantic statistics to drive adaptive ICP selection and raycasting-based dynamic filtering.","core_discovery":"SOCC-ICP jointly executes semantic occupancy grid mapping and LiDAR scan alignment by letting each voxel encode geometric and semantic statistics; this representation supports adaptive selection of point-to-point or point-to-plane ICP, filters dynamic objects via raycasting, and directly supplies a map usable for robotic planning, achieving competitive accuracy that improves further when semantic cues are incorporated.","pith_inferences":["Robotic software stacks could drop one layer of map synchronization code.","The same voxel structure might support incremental semantic segmentation updates across multiple sensors.","Long-term operation in crowded scenes would test whether the raycasting filter accumulates enough free-space evidence to keep drift low."],"forward_implications":["The resulting occupancy grid can be passed directly to motion planners without additional map conversion.","Performance remains competitive in geometrically degenerate environments even when semantic labels are absent.","Adding semantic labels improves accuracy through better downsampling and correspondence weighting.","A single map representation removes the need to maintain separate point-cloud or surfel structures for odometry."],"fun_headline_variants":["SOCC-ICP unifies LiDAR odometry with semantic occupancy grids","Occupancy grid voxels store semantics for adaptive ICP odometry","Unified voxel map for semantics and LiDAR scan alignment","SOCC-ICP uses voxels for both semantic mapping and ICP odometry"],"cache_read_input_tokens":2112,"weakest_assumption_plain":"The method assumes raycasting free-space updates will reliably remove dynamic objects and that local planarity statistics will correctly choose between ICP variants without causing alignment errors in mixed scenes.","fun_headline_variants_meta":{"raw":{"variants":["SOCC-ICP unifies LiDAR odometry with semantic occupancy grids","Occupancy grid voxels store semantics for adaptive ICP odometry","Unified voxel map for semantics and LiDAR scan alignment","SOCC-ICP uses voxels for both semantic mapping and ICP odometry"]},"model":"grok-4.3","cost_usd":0.004429,"raw_usage":{"total_tokens":2121,"prompt_tokens":645,"num_sources_used":0,"completion_tokens":70,"cost_in_usd_ticks":44290500,"prompt_tokens_details":{"text_tokens":645,"audio_tokens":0,"image_tokens":0,"cached_tokens":64},"completion_tokens_details":{"audio_tokens":0,"reasoning_tokens":1406,"accepted_prediction_tokens":0,"rejected_prediction_tokens":0}},"tokens_in":645,"tokens_out":70,"duration_ms":10515,"temperature":1.0,"reasoning_tokens":1406,"cache_read_input_tokens":64,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-06-30T20:12:04.094415+00:00","model_set":{"reader":"grok-4.3"},"falsifier":"Odometry error rising above baseline methods on a sequence containing many independently moving objects that raycasting fails to clear from the grid.","supporting_citations":[],"review_version":1}