[{"data":1,"prerenderedAt":760},["ShallowReactive",2],{"method-balm2_2023":3},{"method":4,"reference":66,"equipment":86,"figures":147,"results":148},{"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":24,"limitations":31,"sensors":40,"platform":42,"estimator":46,"association":47,"timeModel":48,"deskew":49,"loopClosure":50,"globalOptimization":51,"mapRepresentation":52,"prior":53,"outputGeometry":54,"compute":55,"codeUrl":56,"codeLicense":57,"relatedVersions":58},"balm2_2023","Liu et al., 2023a","BALM2 (BALM 2.0)","Efficient and Consistent Bundle Adjustment on Lidar Point Clouds",2023,"recent","C06","offline_map_refinement","BALM2 延續以點到平面或邊緣之歐氏距離為殘差的光達 BA，並提出「點簇（point cluster）」概念，把同一特徵上的所有原始點壓縮為一組緊湊參數，使代價、導數與不確定度計算都不需逐點列舉。作者推導封閉形式的 Jacobian 與 Hessian、其零空間與稀疏性，據此建立二階求解器，並利用二階資訊估計由量測雜訊造成的位姿不確定度。論文另示範其用於光達慣性里程計、多光達外參校正與全域建圖。","BALM2 introduces point clusters that summarize all raw points on a feature, enabling a closed-form second-order BA solver over poses that also estimates pose covariance, without enumerating individual points.","full_text_reviewed","peer_reviewed_published","main_body","以 Hilti 2021 公開資料集評估，其中含 Construction2 工地序列（手持 Ouster OS0-64，軌跡真值來自全測站或動作擷取系統；Sec. VI-B）。在共同初始軌跡下，Construction2 的預設 BALM2 ATE 為 0.0553 m，BALM 為 0.068 m，EF 為 0.086 m（Table II）。此屬公開資料層級的工地證據，並非作者自行於工地部署；地圖品質以無參考之佔用格數（0.1 m 格）評估，未與 TLS 或構件尺寸比對。UzhArea2 牆面點以 BA 位姿配準後平面標準差為 1.7 cm，作者視為達到光達測距雜訊等級（Sec. VI-B1、Fig. 10）。",[20,21,22,23],"simulation","public_benchmark","real_construction_site","independent_reference",[25,26,27,28,29,30],"Avoids per-point enumeration in cost, derivative and uncertainty evaluation (abstract; Sec. VIII-A)","Estimates pose uncertainty from measurement noise; normalized NEES stayed close to one for point noise up to 0.3 m over 2000 Monte Carlo runs (abstract; Sec. III-F; Sec. V, Fig. 7)","Directly minimizes point-to-feature residuals, which the authors argue reinforces map consistency more directly than pose-graph optimization (Sec. VIII-B)","Default double-precision solver had a lower ATE than every non-BALM2 method on all 19 Hilti, VIRAL and UrbanLoco sequences (its float or edge variants were lower still on a few sequences), average 0.1763 m versus 0.221 m for BALM and 0.268 m for EF, and fewer occupied map cells than every non-BALM2 method (Ours (edge) was lower on Campus2) (Sec. VI-B1, VI-B2, Tables II-III)","On Hilti UzhArea2, wall points registered with BA-optimized poses formed a plane with 1.7 cm standard deviation versus 15.3 cm with the provided ground-truth poses (Sec. VI-B1, Fig. 10)","Global BA after MULLS pose-graph output lowered KITTI mean ATE from 1.63 m to 1.34 m (Supplementary I-C, Table VII)",[32,33,34,35,36,37,38,39],"All scans are optimized at once; authors downsampled 10 Hz data to 2 Hz because 10 Hz was prohibitively costly for all BA methods (Sec. VI-B)","Higher computation cost than pose-graph optimization; positioned as accuracy refinement from a baseline trajectory (Sec. VIII-B)","Covariance consistency degrades beyond about 0.3 m point noise, where the first-order approximation no longer holds (Sec. V)","Adding edge features gave no accuracy gain on real data because lidar edge points are noisy (Sec. VI-B1)","Merging planes to cut feature count reduces time but raises pose RMSE because slightly curved surfaces get merged (Sec. VI-B4, Fig. 13)","Motion compensation, dynamic-object removal, tighter fusion with IMU or camera and loop closure are left to future integration (Sec. IX)","Code README warns that large initial pose errors can leave too few detected planes for association and suggests coarse-to-fine re-association (github README)","(inference) Covariance reflects measurement noise under correct association; calibration, timing or association errors are not represented",[41],"3D LiDAR",[43,44,45,20],"handheld","UAV","vehicle","second-order (Newton\u002FLM-type) batch solver over poses after closed-form elimination of plane\u002Fedge parameters; point-cluster coordinates aggregate raw points; LDLT solve (Eigen); pose covariance from second-order information","raw points to plane or edge features via BALM adaptive voxelization of all points registered with an initial trajectory (root voxel 1 m for Hilti, 2 m for VIRAL and UrbanLoco, at most 3 layers, at least 20 points and threshold 1\u002F25 for the feature test); the incremental ICP trajectory served as the common initialization (Sec. VI-B)","discrete poses (authors discuss possible extension to spline or Gaussian-process trajectories, Sec. VIII-C)","no in-scan motion model inside the BA; real-world scans were deskewed by FAST-LIO2 before BA and its odometry output discarded (Sec. VI-B); the simulation ignored in-frame distortion (Sec. V); continuous-time trajectories are only discussed as an extension (Sec. VIII-C)","none within the method; relies on the supplied initial trajectory","batch multi-view BA over all supplied scans","point clusters per plane\u002Fedge feature inside adaptive voxels","initial pose trajectory","optimized scan poses with estimated pose covariance; registered point cloud","offline batch; C++ on Ubuntu 20.04, desktop with Intel i7-10750H CPU and 16 GB RAM; LDLT solve with Eigen 3.3.7, termination at 50 iterations or updates below 1e-6 rad and 1e-6 m (Sec. IV; Sec. VI restates the Sec. IV condition with a 200-iteration limit, an internal inconsistency of the paper); solver complexity O(MfMp + MfMp^2 + Mp^3), independent of the number of points (Sec. III-E, Supplementary IV); real-world scans downsampled from 10 Hz to 2 Hz (Sec. VI-B); on the 19 real sequences the double-precision solver needed about one fourth of BAREG's, one sixth of PA's, one eighth of BALM's and one twentieth of EF's optimization time, with single precision a further 40% faster (Sec. VI-B3, Table IV); as a sliding-window LIO back-end the local BA averaged 73 ms per optimization, keeping 10 Hz (Supplementary I-A, Table V)","https:\u002F\u002Fgithub.com\u002Fhku-mars\u002FBALM","GPL-2.0 (LICENSE file)",[59,63],{"relation":60,"title":61,"doi_or_url":62},"preprint","arXiv 2209.08854 (v2, 2024-06-16)","https:\u002F\u002Farxiv.org\u002Fabs\u002F2209.08854",{"relation":64,"title":65,"doi_or_url":56},"code_release","BALM 2.0 in hku-mars\u002FBALM",{"id":5,"kind":67,"shortName":7,"title":8,"authors":68,"year":9,"venue":72,"venueType":73,"publisher":74,"volumeIssuePages":75,"doi":76,"arxivId":77,"url":62,"firstPublicDate":78,"publicationStatus":16,"metadataStatus":79,"fulltextStatus":15,"era":10,"classicReason":80,"codeUrl":56,"cluster":11,"topics":81,"mdpi":82,"verification":83,"label":6,"fulltextRoute":84,"versionRead":85,"addedByCensus":82},"method",[69,70,71],"Zheng Liu","Xiyuan Liu","Fu Zhang","IEEE Transactions on Robotics","journal","IEEE","39(6):4366-4386","10.1109\u002Ftro.2023.3311671","2209.08854","2022-09-19","metadata_verified","not_applicable",[11],false,"confirmed","arXiv","arXiv v2 (2024-06-16), post-acceptance manuscript of IEEE T-RO 39(6):4366-4386 bundled with the supplementary material; IEEE version of record not read",[87,94,99,103,108,111,116,120,126,131,135,140],{"category":88,"model":89,"canonical":89,"role":90,"dataset":91,"specs":92,"locator":93},"lidar","Ouster OS0-64","dataset sensor","Hilti SLAM Challenge 2021","handheld; lidar data of the Hilti dataset","Sec. VI-B",{"category":95,"model":96,"canonical":96,"role":97,"dataset":91,"specs":98,"locator":93},"total_station","total station (model not reported)","reference or ground truth","ground-truth trajectory for some Hilti sequences",{"category":100,"model":101,"canonical":101,"role":97,"dataset":91,"specs":102,"locator":93},"other","motion capture system (model not reported)","ground-truth trajectory for other Hilti sequences",{"category":88,"model":104,"canonical":105,"role":90,"dataset":106,"specs":107,"locator":93},"OS1 16-channel (two units)","Ouster OS1-16","NTU VIRAL","one horizontal and one vertical on a UAV; only the horizontal one used",{"category":95,"model":109,"canonical":109,"role":97,"dataset":106,"specs":110,"locator":93},"Leica Nova MS60 MultiStation","tracks a crystal prism on the UAV for ground-truth positions",{"category":88,"model":112,"canonical":113,"role":90,"dataset":114,"specs":115,"locator":93},"Velodyne HDL 32E","Velodyne HDL-32E","UrbanLoco","car driving on urban streets",{"category":117,"model":118,"canonical":118,"role":97,"dataset":114,"specs":119,"locator":93},"gnss","Novatel SPAN-CPT","navigation system combining RTK and precise IMU; ground truth (authors note false sudden jumps)",{"category":88,"model":121,"canonical":121,"role":122,"dataset":123,"specs":124,"locator":125},"MID-100","method input","multi-lidar calibration data of [8] and [9]","three internal lidars L0, L1, L2 with 8.4 deg FoV overlap between adjacent ones; in-factory extrinsics used as ground truth","Supplementary I-B, Fig. 13",{"category":88,"model":127,"canonical":128,"role":122,"dataset":129,"specs":130,"locator":125},"AVIA","Livox Avia","multi-lidar calibration data of [8]","lidar L3 without FoV overlap with MID-100",{"category":132,"model":133,"canonical":133,"role":122,"dataset":129,"specs":134,"locator":125},"platform","customized multi-sensor vehicle platform of [8]","rotated one full cycle to create co-visible features",{"category":88,"model":136,"canonical":136,"role":122,"dataset":137,"specs":138,"locator":139},"16-channel lidar (simulated)","authors' simulation","28,800 points per scan, 100 scans along a 92 m rectangular trajectory in a 30 m x 20 m x 8 m semi-closed space","Sec. V, Fig. 5",{"category":141,"model":142,"canonical":142,"role":143,"dataset":144,"specs":145,"locator":146},"compute","Intel i7-10750H","compute for runtime",null,"desktop; 16 GB RAM (written '16Gb RAM'); Ubuntu 20.04","Sec. IV",[],{"totalRows":149,"groupCount":150,"groups":151,"others":743},64,7,[152,487,604,664],{"slug":153,"group":154,"sourceId":5,"sourceLabel":6,"table":155,"selfRows":156,"metrics":157,"seqs":165,"entrants":211,"cells":239,"outcomes":481,"locators":482,"hardware":483,"wordings":484,"notes":485},"balm2-2023-table-ii","balm2_2023:Table II","Table II",34,[158,163],{"label":159,"unit":160,"statistic":161,"alignment":162},"Absolute trajectory error (RMSE, meters)","m","RMSE","not_reported",{"label":164,"unit":160,"statistic":161,"alignment":162},"Absolute trajectory error (RMSE, meters), average over 19 sequences",[166,169,171,173,175,177,179,183,186,188,190,192,194,196,198,200,202,205,207,209],{"dataset":91,"sequence":167,"environment":168},"Basement1","handheld (Ouster OS0-64), Hilti indoor or outdoor sequence",{"dataset":91,"sequence":170,"environment":168},"Basement4",{"dataset":91,"sequence":172,"environment":168},"Campus2",{"dataset":91,"sequence":174,"environment":168},"Construction2",{"dataset":91,"sequence":176,"environment":168},"LabSurvey2",{"dataset":91,"sequence":178,"environment":168},"UzhArea2",{"dataset":180,"sequence":181,"environment":182},"Hilti 2021, VIRAL and UrbanLoco (19 sequences)","Average","mixed: handheld, UAV, car",{"dataset":106,"sequence":184,"environment":185},"eee01","UAV (horizontal 16-channel OS1)",{"dataset":106,"sequence":187,"environment":185},"eee02",{"dataset":106,"sequence":189,"environment":185},"eee03",{"dataset":106,"sequence":191,"environment":185},"nya01",{"dataset":106,"sequence":193,"environment":185},"nya02",{"dataset":106,"sequence":195,"environment":185},"nya03",{"dataset":106,"sequence":197,"environment":185},"sbs01",{"dataset":106,"sequence":199,"environment":185},"sbs02",{"dataset":106,"sequence":201,"environment":185},"sbs03",{"dataset":114,"sequence":203,"environment":204},"0117","car on urban streets (Velodyne HDL 32E)",{"dataset":114,"sequence":206,"environment":204},"0317",{"dataset":114,"sequence":208,"environment":204},"0426-1",{"dataset":114,"sequence":210,"environment":204},"0426-2",[212,216,219,221,224,227,229,231,233,235,237],{"name":213,"methodId":214,"linkable":215,"proposed":82,"self":82},"ICP (PCL, incremental)","besl1992icp",true,{"name":217,"methodId":218,"linkable":215,"proposed":82,"self":82},"GICP (PCL, incremental)","segal2009gicp",{"name":220,"methodId":144,"linkable":82,"proposed":82,"self":82},"NDT (PCL, incremental)",{"name":222,"methodId":223,"linkable":215,"proposed":82,"self":82},"EF","eigenfactors2019",{"name":225,"methodId":226,"linkable":215,"proposed":82,"self":82},"BALM","balm2021",{"name":228,"methodId":144,"linkable":82,"proposed":82,"self":82},"PA",{"name":230,"methodId":144,"linkable":82,"proposed":82,"self":82},"PA (inner)",{"name":232,"methodId":144,"linkable":82,"proposed":82,"self":82},"BAREG",{"name":234,"methodId":5,"linkable":215,"proposed":215,"self":215},"Ours (float)",{"name":236,"methodId":5,"linkable":215,"proposed":215,"self":215},"Ours (edge)",{"name":238,"methodId":5,"linkable":215,"proposed":215,"self":215},"Ours",[240,244,247,250,253,256,259,262,264,267,270,273,275,277,279,281,282,284,286,288,290,292,294,296,298,300,302,304,305,306,307,309,311,313,315,317,319,321,323,325,327,328,330,332,334,335,337,339,341,343,345,346,348,350,352,354,356,358,360,362,364,366,368,370,372,374,376,378,380,382,384,386,388,390,392,394,396,398,400,402,404,406,408,410,411,413,415,417,419,421,424,426,428,431,433,435,438,440,442,445,446,448,450,451,453,456,458,460,463,465,467,470,472,474,477,479],[241,241,241,242,243,241,243,243,241],0,0.058,-1,[245,241,241,246,243,241,243,243,241],1,0.063,[248,241,241,249,243,241,243,243,241],2,0.076,[251,241,241,252,243,241,243,243,241],3,0.047,[254,241,241,255,243,241,243,243,241],4,0.042,[257,241,241,258,243,241,243,243,241],5,0.038,[260,241,241,261,243,241,243,243,241],6,0.036,[150,241,241,263,243,241,243,243,241],0.04,[265,241,241,266,243,241,243,243,241],8,0.0359,[268,241,241,269,243,241,243,243,241],9,0.0361,[271,241,241,272,243,241,243,243,241],10,0.0353,[241,241,245,274,243,241,243,243,241],0.084,[245,241,245,276,243,241,243,243,241],0.089,[248,241,245,278,243,241,243,243,241],0.098,[251,241,245,280,243,241,243,243,241],0.071,[254,241,245,242,243,241,243,243,241],[257,241,245,283,243,241,243,243,241],0.048,[260,241,245,285,243,241,243,243,241],0.045,[150,241,245,287,243,241,243,243,241],0.054,[265,241,245,289,243,241,243,243,241],0.0444,[268,241,245,291,243,241,243,243,241],0.0448,[271,241,245,293,243,241,243,243,241],0.0443,[241,241,248,295,243,241,243,243,241],0.105,[245,241,248,297,243,241,243,243,241],0.109,[248,241,248,299,243,241,243,243,241],0.124,[251,241,248,301,243,241,243,243,241],0.08,[254,241,248,303,243,241,243,243,241],0.066,[257,241,248,242,243,241,243,243,241],[260,241,248,287,243,241,243,243,241],[150,241,248,246,243,241,243,243,241],[265,241,248,308,243,241,243,243,241],0.0535,[268,241,248,310,243,241,243,243,241],0.053,[271,241,248,312,243,241,243,243,241],0.0531,[241,241,251,314,243,241,243,243,241],0.108,[245,241,251,316,243,241,243,243,241],0.104,[248,241,251,318,243,241,243,243,241],0.113,[251,241,251,320,243,241,243,243,241],0.086,[254,241,251,322,243,241,243,243,241],0.068,[257,241,251,324,243,241,243,243,241],0.06,[260,241,251,326,243,241,243,243,241],0.059,[150,241,251,246,243,241,243,243,241],[265,241,251,329,243,241,243,243,241],0.0563,[268,241,251,331,243,241,243,243,241],0.0577,[271,241,251,333,243,241,243,243,241],0.0553,[241,241,254,303,243,241,243,243,241],[245,241,254,336,243,241,243,243,241],0.069,[248,241,254,338,243,241,243,243,241],0.072,[251,241,254,340,243,241,243,243,241],0.046,[254,241,254,342,243,241,243,243,241],0.025,[257,241,254,344,243,241,243,243,241],0.019,[260,241,254,344,243,241,243,243,241],[150,241,254,347,243,241,243,243,241],0.023,[265,241,254,349,243,241,243,243,241],0.0185,[268,241,254,351,243,241,243,243,241],0.0189,[271,241,254,353,243,241,243,243,241],0.0181,[241,241,257,355,243,241,243,243,241],0.182,[245,241,257,357,243,241,243,243,241],0.191,[248,241,257,359,243,241,243,243,241],0.211,[251,241,257,361,243,241,243,243,241],0.161,[254,241,257,363,243,241,243,243,241],0.141,[257,241,257,365,243,241,243,243,241],0.122,[260,241,257,367,243,241,243,243,241],0.121,[150,241,257,369,243,241,243,243,241],0.127,[265,241,257,371,243,241,243,243,241],0.1205,[268,241,257,373,243,241,243,243,241],0.1102,[271,241,257,375,243,241,243,243,241],0.1171,[241,245,260,377,243,241,243,243,241],0.411,[245,245,260,379,243,241,243,243,241],0.41,[248,245,260,381,243,241,243,243,241],0.412,[251,245,260,383,243,241,243,243,241],0.268,[254,245,260,385,243,241,243,243,241],0.221,[257,245,260,387,243,241,243,243,241],0.186,[260,245,260,389,243,241,243,243,241],0.179,[150,245,260,391,243,241,243,243,241],0.203,[265,245,260,393,243,241,243,243,241],0.1775,[268,245,260,395,243,241,243,243,241],0.1826,[271,245,260,397,243,241,243,243,241],0.1763,[251,241,150,399,243,241,243,243,241],0.102,[254,241,150,401,243,241,243,243,241],0.073,[271,241,150,403,243,241,243,243,241],0.0382,[251,241,265,405,243,241,243,243,241],0.092,[254,241,265,407,243,241,243,243,241],0.062,[271,241,265,409,243,241,243,243,241],0.0356,[251,241,268,318,243,241,243,243,241],[254,241,268,412,243,241,243,243,241],0.081,[271,241,268,414,243,241,243,243,241],0.0517,[251,241,271,416,243,241,243,243,241],0.107,[254,241,271,418,243,241,243,243,241],0.082,[271,241,271,420,243,241,243,243,241],0.0362,[251,241,422,423,243,241,243,243,241],11,0.097,[254,241,422,425,243,241,243,243,241],0.067,[271,241,422,427,243,241,243,243,241],0.0468,[251,241,429,430,243,241,243,243,241],12,0.085,[254,241,429,432,243,241,243,243,241],0.074,[271,241,429,434,243,241,243,243,241],0.0413,[251,241,436,437,243,241,243,243,241],13,0.083,[254,241,436,439,243,241,243,243,241],0.077,[271,241,436,441,243,241,243,243,241],0.0385,[251,241,443,444,243,241,243,243,241],14,0.094,[254,241,443,407,243,241,243,243,241],[271,241,443,447,243,241,243,243,241],0.0377,[251,241,449,314,243,241,243,243,241],15,[254,241,449,338,243,241,243,243,241],[271,241,449,452,243,241,243,243,241],0.0427,[251,241,454,455,243,241,243,243,241],16,0.728,[254,241,454,457,243,241,243,243,241],0.625,[271,241,454,459,243,241,243,243,241],0.4956,[251,241,461,462,243,241,243,243,241],17,0.878,[254,241,461,464,243,241,243,243,241],0.732,[271,241,461,466,243,241,243,243,241],0.6488,[251,241,468,469,243,241,243,243,241],18,1.014,[254,241,468,471,243,241,243,243,241],0.875,[271,241,468,473,243,241,243,243,241],0.6886,[251,241,475,476,243,241,243,243,241],19,1.113,[254,241,475,478,243,241,243,243,241],0.924,[271,241,475,480,243,241,243,243,241],0.8223,[],[155],[],[],[486],"ATE RMSE (m) of multi-view registration; scans deskewed by FAST-LIO2 (odometry output discarded) and downsampled from 10 Hz to 2 Hz; ICP, GICP, NDT from PCL run incrementally against the last 20 scans; the ICP trajectory is the common initialization and adaptive voxelization (root voxel 1 m Hilti, 2 m VIRAL and UrbanLoco) the common association for EF, BALM, PA and Ours (BAREG uses its own); plane features only except Ours (edge); ground truth: Hilti total station or motion capture, VIRAL Leica Nova MS60, UrbanLoco Novatel SPAN-CPT RTK\u002FINS. Column order verified from PDF layout: Ours (float), Ours (edge), Ours",{"slug":488,"group":489,"sourceId":490,"sourceLabel":491,"table":492,"selfRows":429,"metrics":493,"seqs":504,"entrants":518,"cells":526,"outcomes":598,"locators":599,"hardware":600,"wordings":601,"notes":602},"lemon2026-table-iii","lemon2026:Table III","lemon2026","Wang et al., 2026","Table III",[494,499,502],{"label":495,"unit":496,"statistic":497,"alignment":498},"MME","unitless","mean","none",{"label":500,"unit":162,"statistic":497,"alignment":501},"z-Drift","first-pose",{"label":503,"unit":162,"statistic":161,"alignment":501},"z-RMSE",[505,509,512,515],{"dataset":506,"sequence":507,"environment":508},"self-collected (Mid360)","Garage","indoor, 334 m",{"dataset":506,"sequence":510,"environment":511},"Library","outdoor, 519 m",{"dataset":506,"sequence":513,"environment":514},"Yard","indoor and outdoor, 232 m",{"dataset":506,"sequence":516,"environment":517},"Laboratory","outdoor, 98 m",[519,522,524],{"name":520,"methodId":521,"linkable":215,"proposed":82,"self":82},"HBA","hba2023",{"name":523,"methodId":5,"linkable":215,"proposed":82,"self":215},"BALM2",{"name":525,"methodId":490,"linkable":215,"proposed":215,"self":82},"Ours (spatial BA)",[527,529,531,533,535,537,539,541,543,545,547,549,551,553,555,557,559,561,563,565,567,569,571,573,575,577,579,581,583,585,587,589,591,593,594,596],[241,241,241,528,243,241,243,243,241],-6.83,[241,245,241,530,243,241,243,243,241],5.98,[241,248,241,532,243,241,243,243,241],7.14,[245,241,241,534,243,241,243,243,241],-6.93,[245,245,241,536,243,241,243,243,241],5.95,[245,248,241,538,243,241,243,243,241],7.17,[248,241,241,540,243,241,243,243,241],-6.86,[248,245,241,542,243,241,243,243,241],4.87,[248,248,241,544,243,241,243,243,241],5.35,[241,241,245,546,243,241,243,243,241],-5.95,[241,245,245,548,243,241,243,243,241],6.17,[241,248,245,550,243,241,243,243,241],6.78,[245,241,245,552,243,241,243,243,241],-6.2,[245,245,245,554,243,241,243,243,241],6.51,[245,248,245,556,243,241,243,243,241],7.31,[248,241,245,558,243,241,243,243,241],-6.23,[248,245,245,560,243,241,243,243,241],3.68,[248,248,245,562,243,241,243,243,241],4.53,[241,241,248,564,243,241,243,243,241],-6.39,[241,245,248,566,243,241,243,243,241],2.33,[241,248,248,568,243,241,243,243,241],2.7,[245,241,248,570,243,241,243,243,241],-5.93,[245,245,248,572,243,241,243,243,241],1.86,[245,248,248,574,243,241,243,243,241],2.12,[248,241,248,576,243,241,243,243,241],-6.4,[248,245,248,578,243,241,243,243,241],1.12,[248,248,248,580,243,241,243,243,241],1.25,[241,241,251,582,243,241,243,243,241],-6.26,[241,245,251,584,243,241,243,243,241],3.24,[241,248,251,586,243,241,243,243,241],3.72,[245,241,251,588,243,241,243,243,241],-6.04,[245,245,251,590,243,241,243,243,241],3.22,[245,248,251,592,243,241,243,243,241],3.7,[248,241,251,552,243,241,243,243,241],[248,245,251,595,243,241,243,243,241],3.17,[248,248,251,597,243,241,243,243,241],3.64,[],[492],[],[],[603],"Single-robot study on self-collected Mid360 data; spatial BA versus BALM2 (sliding window) and HBA, all on raw odometry without loop-based refinement; z-drift and z-RMSE relative to the z-value of the first frame; MME via MapEval (lower is better)",{"slug":605,"group":606,"sourceId":5,"sourceLabel":6,"table":607,"selfRows":260,"metrics":608,"seqs":611,"entrants":615,"cells":624,"outcomes":657,"locators":658,"hardware":659,"wordings":661,"notes":662},"balm2-2023-table-iv","balm2_2023:Table IV","Table IV",[609],{"label":610,"unit":162,"statistic":162,"alignment":498},"Optimization time (total, unit not stated)",[612,614],{"dataset":91,"sequence":174,"environment":613},"handheld (Ouster OS0-64), construction sequence",{"dataset":180,"sequence":181,"environment":182},[616,617,618,619,620,621,622,623],{"name":222,"methodId":223,"linkable":215,"proposed":82,"self":82},{"name":225,"methodId":226,"linkable":215,"proposed":82,"self":82},{"name":228,"methodId":144,"linkable":82,"proposed":82,"self":82},{"name":230,"methodId":144,"linkable":82,"proposed":82,"self":82},{"name":232,"methodId":144,"linkable":82,"proposed":82,"self":82},{"name":234,"methodId":5,"linkable":215,"proposed":215,"self":215},{"name":236,"methodId":5,"linkable":215,"proposed":215,"self":215},{"name":238,"methodId":5,"linkable":215,"proposed":215,"self":215},[625,627,629,631,633,635,637,639,641,643,645,647,649,651,653,655],[241,241,241,626,243,241,241,243,241],1415.18,[245,241,241,628,243,241,241,243,241],412,[248,241,241,630,243,241,241,243,241],335.7,[251,241,241,632,243,241,241,243,241],313.23,[254,241,241,634,243,241,241,243,241],231.48,[257,241,241,636,243,241,241,243,241],33.04,[260,241,241,638,243,241,241,243,241],47.34,[150,241,241,640,243,241,241,243,241],47.12,[241,241,245,642,243,241,241,243,241],647.29,[245,241,245,644,243,241,241,243,241],232.54,[248,241,245,646,243,241,241,243,241],202.64,[251,241,245,648,243,241,241,243,241],171.11,[254,241,245,650,243,241,241,243,241],132.1,[257,241,245,652,243,241,241,243,241],18.15,[260,241,245,654,243,241,241,243,241],31.16,[150,241,245,656,243,241,241,243,241],30.58,[],[607],[660],"desktop Intel i7-10750H, 16 GB RAM",[],[663],"Total optimization time of the BA methods on the Table II inputs (pairwise methods excluded); the table does not state the time unit; desktop Intel i7-10750H, 16 GB RAM (Sec. IV)",{"slug":665,"group":666,"sourceId":521,"sourceLabel":667,"table":668,"selfRows":260,"metrics":669,"seqs":675,"entrants":684,"cells":694,"outcomes":736,"locators":737,"hardware":739,"wordings":740,"notes":741},"hba2023-table-viii","hba2023:Table VIII","Liu et al., 2023b","Table VIII",[670,672],{"label":671,"unit":160,"statistic":161,"alignment":162},"RMSE of ATE",{"label":673,"unit":674,"statistic":162,"alignment":498},"total optimization time","s",[676,680,682],{"dataset":677,"sequence":678,"environment":679},"MulRan","DCC01","urban campus driving",{"dataset":677,"sequence":681,"environment":679},"DCC02",{"dataset":677,"sequence":683,"environment":679},"DCC03",[685,688,690,692],{"name":686,"methodId":687,"linkable":215,"proposed":82,"self":82},"Initial (LIO-SAM with loop closure)","liosam2020",{"name":689,"methodId":5,"linkable":215,"proposed":82,"self":215},"Original BA",{"name":691,"methodId":144,"linkable":82,"proposed":82,"self":82},"Reduced BA",{"name":693,"methodId":521,"linkable":215,"proposed":215,"self":82},"Proposed",[695,697,699,701,703,705,707,709,711,713,715,717,719,721,723,725,727,729,731,733,734],[241,241,241,696,243,241,243,243,241],5.67,[241,241,245,698,243,241,243,243,241],3.48,[241,241,248,700,243,241,243,243,241],2.84,[245,241,241,702,243,241,243,243,241],5.17,[245,245,241,704,243,241,243,243,241],2830.91,[245,241,245,706,243,241,243,243,241],3.19,[245,245,245,708,243,241,243,243,241],2631.83,[245,241,248,710,243,241,243,243,241],2.54,[245,245,248,712,243,241,243,243,241],3655.1,[248,241,241,714,243,241,243,243,241],5.66,[248,245,241,716,243,241,243,243,241],4390.15,[248,241,245,718,243,241,243,243,241],3.46,[248,245,245,720,243,241,243,243,241],4615.9,[248,241,248,722,243,241,243,243,241],2.83,[248,245,248,724,243,241,243,243,241],7522.2,[251,241,241,726,243,241,243,243,241],5.19,[251,245,241,728,243,241,243,243,241],226.1,[251,241,245,730,243,241,243,243,241],3.2,[251,245,245,732,243,241,243,243,241],362.59,[251,241,248,710,243,241,243,243,241],[251,245,248,735,243,241,243,243,241],248.42,[],[738],"Table VIII (version of record)",[],[],[742],"MulRan DCC sequences with LIO-SAM loop-closed input ('Initial'); original BA, reduced block-diagonal BA and proposed hierarchical BA; RMSE of ATE and total optimization time; hardware not reported",[744,748,754],{"group":745,"slug":746,"sourceLabel":6,"table":492,"selfRows":251,"datasets":747},"balm2_2023:Table III","balm2-2023-table-iii",[91],{"group":749,"slug":750,"sourceLabel":6,"table":751,"selfRows":248,"datasets":752},"balm2_2023:Supplementary Table V","balm2-2023-supplementary-table-v","Supplementary Table V",[753],"utbm, uclk, nclt (11 sequences)",{"group":755,"slug":756,"sourceLabel":6,"table":757,"selfRows":245,"datasets":758},"balm2_2023:Supplementary Table VII","balm2-2023-supplementary-table-vii","Supplementary Table VII",[759],"KITTI odometry",1790510654894]