[{"data":1,"prerenderedAt":250},["ShallowReactive",2],{"method-ebadi2021dareslam":3},{"method":4,"reference":52,"equipment":74,"figures":111,"results":149},{"id":5,"label":6,"shortName":7,"title":8,"year":9,"era":10,"cluster":11,"scope":12,"keyIdeaZh":13,"keyIdeaEn":14,"fulltextStatus":15,"publicationStatus":16,"recommendation":17,"constructionRelevance":18,"validationEnvironment":19,"strengths":22,"limitations":27,"sensors":32,"platform":35,"estimator":37,"association":38,"timeModel":39,"deskew":39,"loopClosure":40,"globalOptimization":41,"mapRepresentation":42,"prior":43,"outputGeometry":44,"compute":45,"codeUrl":46,"codeLicense":47,"relatedVersions":48},"ebadi2021dareslam","Ebadi et al., 2021","DARE-SLAM","DARE-SLAM: Degeneracy-Aware and Resilient Loop Closing in Perceptually-Degraded Environments",2021,"recent","C13","full_slam_with_global_correction","DARE-SLAM 先以 ICP 解的特徵分析估計環境的幾何退化程度，並把模糊、不可觀測的區域排除在迴圈閉合搜尋之外，以免錯誤迴圈扭曲整張地圖。再以 LiDAR 點雲的 2D 與 3D 顯著特徵進行對漂移較不敏感的迴圈閉合。作者在礦坑與辦公室資料上評估。","Uses ICP eigen-analysis to gate loop-closure search away from degenerate regions and adds a drift-resilient loop closure based on salient LiDAR features.","full_text_reviewed","peer_reviewed_published","background","未在工地測試；地下礦坑的長直無特徵廊道結論可類比隧道施工（推論）。",[20,21],"underground_or_tunnel","completed_building",[23,24,25,26],"SGLC loop closing gave lower APE than LeGO-LOAM and BGLC in four subterranean datasets (10 runs each; APE from known landmark locations used as proxy ground truth) (Sec. 4, Fig. 25)","Geometric degeneracy detector reached AUC 0.887 on 254 manually labelled scans (61 degenerate) (Sec. 3.2, Fig. 7)","Occupancy-grid pre-matching reached an average AUC of 0.756 across five environments (Sec. 4.1, Fig. 17)","Scan-to-submap refinement cut average drift per 300 m from 6% to below 1% (Safety Research mine) and from 27% to 2% (Experimental mine) (Sec. 3.1)",[28,29,30,31],"All compared methods, including the proposed one, showed larger drift in mines with long, flat, featureless corridors (Sec. 4, discussion of Fig. 25)","Pre-matching produces more false positives in self-similar indoor offices (identical cubicles) than in mines, so geometric verification and outlier rejection are still required (Sec. 4.1)","Even with scan-to-submap refinement the front-end drift in the Experimental mine remains substantially large, so loop closure is still needed (Sec. 3.1)","Ground truth is a proxy built from known object and fiducial locations, not a full reference trajectory (Sec. 4)",[33,34],"3D LiDAR (Velodyne VLP-16 Puck Lite)","RGB-D camera (RealSense D435, object detection only)",[36],"wheeled UGVs (Husky A200 series; single robot and two-robot teams in six US mines and an indoor office)","LOAM-style edge and planar feature extraction (up to 90% point decimation), then two-stage GICP (scan-to-scan, then scan-to-submap) odometry; reduced pose graph with a key-node every 1 m translation or 30 deg rotation, back-end in GTSAM optimized by iterative nonlinear least squares (Levenberg-Marquardt named only as an example); local graphs per robot merged at a base station","GICP correspondences with approximate nearest-neighbour search against a local submap; degeneracy from the condition number of the ICP-derived approximate Hessian (log kappa threshold); loop-closure pre-matching with ORB features on binary occupancy-grid images, FLANN matching and RANSAC homography, scored by correspondence confidence times transformation confidence; ICP geometric verification seeded with the homography yaw","not_reported","degeneracy-aware candidate gating plus pose-invariant multi-stage loop closing (SGLC): global pre-matching of 250 x 250-cell (5 m x 5 m) occupancy-grid images over the whole trajectory (maps with 20 or fewer inliers dropped), ICP geometric verification, and PCM pairwise-consistency outlier rejection; same pipeline for inter-robot loops","pose graph optimization with a GTSAM back-end (iterative nonlinear optimization, Levenberg-Marquardt named only as an example) after PCM outlier rejection; single robot and centralized multi-robot merging at a base station assuming known initial robot poses","reduced pose graph whose key-nodes carry key-scans; global 3D point-cloud map formed by projecting key-scans into the world frame; per-key-scan 2D occupancy grids used only for place recognition","known initial robot poses in a common world frame for multi-robot merging; proxy ground truth obtained by enforcing the known ground-truth locations of objects and fiducial markers (provided by DARPA) in each robot's pose graph (evaluation only)","3D point-cloud maps of mines (single-robot and merged multi-robot) and optimized robot trajectories","Onboard Intel NUC 7i7DNBE (4 x 1.9 GHz, 32 GB RAM) per robot; base station Intel Hades Canyon NUC8i7HVKVA (4 x 1.9 GHz, 32 GB RAM), for which real time was not strictly required; pre-matching can run on a separate thread; execution times given only as box plots (Fig. 18)",null,"not_verified",[49],{"relation":50,"title":8,"doi_or_url":51},"preprint","https:\u002F\u002Farxiv.org\u002Fabs\u002F2102.05117",{"id":5,"kind":53,"shortName":7,"title":8,"authors":54,"year":9,"venue":60,"venueType":61,"publisher":62,"volumeIssuePages":63,"doi":64,"arxivId":65,"url":51,"firstPublicDate":66,"publicationStatus":16,"metadataStatus":67,"fulltextStatus":15,"era":10,"classicReason":68,"codeUrl":46,"cluster":11,"topics":69,"mdpi":70,"verification":71,"label":6,"fulltextRoute":72,"versionRead":73,"addedByCensus":70},"method",[55,56,57,58,59],"Kamak Ebadi","Matteo Palieri","Sally Wood","Curtis Padgett","Ali-akbar Agha-mohammadi","Journal of Intelligent & Robotic Systems","journal","Springer","102(1), article 2","10.1007\u002Fs10846-021-01362-w","2102.05117","2021-02-09","metadata_verified","not_applicable",[11],false,"corrected","arXiv","arXiv 2102.05117v1 (9 Feb 2021, 33 pp., author manuscript marked 'Accepted for publication in the Journal of Intelligent and Robotic Systems'); Springer version of record (JIRS 102, article 2) not licensed at NTU, only its landing page (abstract, declarations) was read",[75,82,87,92,97,100,106],{"category":76,"model":77,"canonical":78,"role":79,"dataset":46,"specs":80,"locator":81},"lidar","Velodyne VLP-16 Puck Lite","Velodyne VLP-16","method input","not_reported (low vertical resolution noted as a cause of poor scan-to-scan estimates in narrow tunnels)","Sec. 4; Sec. 3.1",{"category":83,"model":84,"canonical":84,"role":79,"dataset":46,"specs":85,"locator":86},"rgbd","Intel RealSense D435","RGB-D camera used for YOLO-based object detection and localization, not for SLAM","Sec. 4",{"category":88,"model":89,"canonical":89,"role":79,"dataset":46,"specs":90,"locator":91},"platform","Husky A200 series","wheeled ground robot; up to two robots in multi-robot trials","Sec. 4, Fig. 16",{"category":93,"model":94,"canonical":94,"role":95,"dataset":46,"specs":96,"locator":86},"compute","Intel NUC 7i7DNBE","compute for runtime","4 x 1.9 GHz, 32 GB RAM; onboard SLAM on each robot",{"category":93,"model":98,"canonical":98,"role":95,"dataset":46,"specs":99,"locator":86},"Intel Hades Canyon NUC8i7HVKVA","4 x 1.9 GHz, 32 GB RAM; base station merging local pose graphs",{"category":101,"model":102,"canonical":102,"role":103,"dataset":46,"specs":104,"locator":105},"wheel_or_leg_odometry","wheel-inertial odometry (sensor models not reported)","reference or ground truth","reference for lidar slip in a carpeted office corridor at low speed; wheel slippage stated to be negligible","Sec. 3.2, Fig. 6",{"category":107,"model":108,"canonical":108,"role":103,"dataset":46,"specs":109,"locator":110},"other","fiducial markers and objects with known ground-truth locations (provided by DARPA)","known locations enforced in each robot's pose graph to build proxy ground-truth trajectories; object positions also used to score object localization error","Sec. 4, Sec. 4.3",[112,125,133,141],{"refId":5,"refLabel":6,"fig":113,"whatZh":114,"license":115,"licenseUrl":116,"sourceUrl":117,"src":118,"width":119,"height":120,"thumb":121,"thumbWidth":122,"thumbHeight":123,"modified":124},"Fig. 3","LiDAR 前端流程（點雲過濾、scan-to-scan 與 scan-to-submap 配準）與局部位姿圖","CC BY 4.0 (arXiv v1 abs page); note: the arXiv PDF title-page footnote also prints '©2020 All rights reserved'","https:\u002F\u002Fcreativecommons.org\u002Flicenses\u002Fby\u002F4.0\u002F","https:\u002F\u002Farxiv.org\u002Fhtml\u002F2102.05117v1\u002Foverview.png","\u002Ffigure-files\u002Febadi2021dareslam\u002Ffig-3.webp",1400,444,"\u002Ffigure-files\u002Febadi2021dareslam\u002Ffig-3.thumb.webp",480,152,"resized to at most 1400 px wide and converted to WebP",{"refId":5,"refLabel":6,"fig":126,"whatZh":127,"license":115,"licenseUrl":116,"sourceUrl":128,"src":129,"width":119,"height":130,"thumb":131,"thumbWidth":122,"thumbHeight":132,"modified":124},"Fig. 10","顯著性迴圈閉合流程：佔據格地圖預匹配、幾何驗證與離群值剔除","https:\u002F\u002Farxiv.org\u002Fhtml\u002F2102.05117v1\u002Floop_closure_diagram.png","\u002Ffigure-files\u002Febadi2021dareslam\u002Ffig-10.webp",950,"\u002Ffigure-files\u002Febadi2021dareslam\u002Ffig-10.thumb.webp",326,{"refId":5,"refLabel":6,"fig":134,"whatZh":135,"license":115,"licenseUrl":116,"sourceUrl":136,"src":137,"width":119,"height":138,"thumb":139,"thumbWidth":122,"thumbHeight":140,"modified":124},"Fig. 21","Eagle 礦坑雙機器人地圖：無迴圈閉合、BGLC 與本法在位姿圖最佳化前後的對照","https:\u002F\u002Farxiv.org\u002Fhtml\u002F2102.05117v1\u002Flc_eagle_mine.png","\u002Ffigure-files\u002Febadi2021dareslam\u002Ffig-21.webp",668,"\u002Ffigure-files\u002Febadi2021dareslam\u002Ffig-21.thumb.webp",229,{"refId":5,"refLabel":6,"fig":142,"whatZh":143,"license":115,"licenseUrl":116,"sourceUrl":144,"src":145,"width":119,"height":146,"thumb":147,"thumbWidth":122,"thumbHeight":148,"modified":124},"Fig. 23","辦公室、Beckley 煤礦與 Safety Research 礦坑以 LeGO-LOAM、BGLC 與本法建立的地圖比較","https:\u002F\u002Farxiv.org\u002Fhtml\u002F2102.05117v1\u002Fmap_comparisons1.png","\u002Ffigure-files\u002Febadi2021dareslam\u002Ffig-23.webp",867,"\u002Ffigure-files\u002Febadi2021dareslam\u002Ffig-23.thumb.webp",297,{"totalRows":150,"groupCount":151,"groups":152,"others":249},4,3,[153,198,224],{"slug":154,"group":155,"sourceId":5,"sourceLabel":6,"table":156,"selfRows":157,"metrics":158,"seqs":163,"entrants":171,"cells":177,"outcomes":187,"locators":192,"hardware":194,"wordings":195,"notes":196},"ebadi2021dareslam-text-sec-3-1","ebadi2021dareslam:Text Sec. 3.1","Text Sec. 3.1",2,[159],{"label":160,"unit":161,"statistic":162,"alignment":39},"average translational drift per 300 m (relative position error)","%","mean",[164,168],{"dataset":165,"sequence":166,"environment":167},"authors' DARPA SubT Tunnel Circuit recordings","Bruceton Safety Research mine (1400 m)","underground coal mine",{"dataset":165,"sequence":169,"environment":170},"Bruceton Experimental mine (700 m)","underground coal mine with long featureless corridors",[172,174],{"name":173,"methodId":46,"linkable":70,"proposed":70,"self":70},"scan-to-scan GICP registration only",{"name":175,"methodId":5,"linkable":176,"proposed":176,"self":176},"scan-to-scan plus scan-to-submap GICP (two-stage front-end from LAMP used in DARE-SLAM)",true,[178,182,184,186],[179,179,179,180,179,179,181,181,179],0,6,-1,[183,179,179,183,183,179,181,181,179],1,[179,179,183,185,157,179,181,181,179],27,[183,179,183,157,151,179,181,181,179],[188,189,190,191],"stated as more than 18 m average drift per 300 m (6%)","stated as less than 1% of distance per 300 m","stated as more than 80 m average drift per 300 m (27%)","stated as 7 m average drift per 300 m (2%)",[193],"Sec. 3.1",[],[],[197],"Front-end odometry drift from EVO relative pose error per 300 m travelled in autonomous traverses; values stated in text (Fig. 4 box plots); percentages are relative position error",{"slug":199,"group":200,"sourceId":5,"sourceLabel":6,"table":201,"selfRows":183,"metrics":202,"seqs":206,"entrants":211,"cells":214,"outcomes":217,"locators":218,"hardware":220,"wordings":221,"notes":222},"ebadi2021dareslam-text-sec-3-2","ebadi2021dareslam:Text Sec. 3.2","Text Sec. 3.2",[203],{"label":204,"unit":205,"statistic":39,"alignment":39},"area under ROC curve (AUC) of the degeneracy detector","unitless",[207],{"dataset":208,"sequence":209,"environment":210},"authors' labelled scans","254 lidar scans","not stated for the ROC set",[212],{"name":213,"methodId":5,"linkable":176,"proposed":176,"self":176},"degeneracy detector (log condition number of approximate Hessian)",[215],[179,179,179,216,181,179,181,181,179],0.887,[],[219],"Sec. 3.2, Fig. 7",[],[],[223],"ROC analysis of the geometric degeneracy detector on 254 manually labelled lidar scans, 61 of them degenerate",{"slug":225,"group":226,"sourceId":5,"sourceLabel":6,"table":227,"selfRows":183,"metrics":228,"seqs":231,"entrants":236,"cells":239,"outcomes":242,"locators":243,"hardware":245,"wordings":246,"notes":247},"ebadi2021dareslam-text-sec-4-1","ebadi2021dareslam:Text Sec. 4.1","Text Sec. 4.1",[229],{"label":230,"unit":205,"statistic":162,"alignment":39},"average AUC for identifying loop closures",[232],{"dataset":233,"sequence":234,"environment":235},"authors' recordings","five environments","described as underground; discussion also covers an indoor office",[237],{"name":238,"methodId":5,"linkable":176,"proposed":176,"self":176},"SGLC pre-matching (similarity confidence)",[240],[179,179,179,241,181,179,181,181,179],0.756,[],[244],"Sec. 4.1, Fig. 17",[],[],[248],"ROC analysis of occupancy-grid pre-matching; 100 salient grid maps per environment, 20 of which are true loop closures",[],1790510661161]