[{"data":1,"prerenderedAt":495},["ShallowReactive",2],{"method-locus2_2022":3},{"method":4,"reference":71,"equipment":97,"figures":139,"results":140},{"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":23,"limitations":29,"sensors":36,"platform":40,"estimator":43,"association":44,"timeModel":45,"deskew":46,"loopClosure":47,"globalOptimization":48,"mapRepresentation":49,"prior":50,"outputGeometry":51,"compute":52,"codeUrl":53,"codeLicense":54,"relatedVersions":55},"locus2_2022","Reinke et al., 2022","LOCUS 2.0","LOCUS 2.0: Robust and Computationally Efficient Lidar Odometry for Real-Time 3D Mapping",2022,"recent","C04","odometry_with_local_mapping","LOCUS 2.0 是以 LiDAR 為核心、可鬆耦合其他里程計的多階段 GICP 里程計，針對算力與記憶體受限的地下探勘機器人設計。它把 GICP 所需的點共變異數改由預先計算的法向量直接構成，地圖點不必重算共變異數；以自適應體素濾波把每幀點數維持在設定值附近，使運算時間不隨環境大小或 LiDAR 數量劇烈變動；地圖只保留以機器人為中心 50 m 的滑動視窗，可用多執行緒八元樹或 ikd-Tree 儲存，以限制記憶體用量。系統本身沒有迴圈閉合。","LiDAR-centric multi-stage GICP odometry for compute- and memory-constrained underground robots: point covariances built from stored normals, an adaptive voxel filter that holds the point count constant, and a 50 m sliding-window map (multi-threaded octree or ikd-tree); optional non-lidar odometry seeds the registration; no loop closure.","full_text_reviewed","peer_reviewed_published","background","未在施工現場測試；資料來自 DARPA SubT 的停用電廠、礦坑隧道、地鐵站與熔岩洞，其中電廠與地鐵站屬既有建築或基礎設施。狹長走廊與隧道的幾何退化和地下工程施工相似，作者報告在隧道資料 F 中 FAST-LIO 與 LINS 誤差極大而 LOCUS 2.0 仍可運作；但真值是以 LOCUS 1.0 對測量等級地圖配準產生，並非完全獨立的量測。另一篇基礎設施檢測回顧的作者表示無法執行 LOCUS 的程式碼 [ghadimzadeh2025slamnde]。",[20,21,22],"underground_or_tunnel","infrastructure","completed_building",[24,25,26,27,28],"Smallest max and mean APE on 4 of the 6 compared underground datasets (A, C, F and H); on I LINS and on J FAST-LIO have lower APE, although the text states 5 of 6 (Table III; Sec. IV-E)","Only compared method that did not fail in the Bruceton Mine tunnel dataset F (Table III; Sec. IV-E)","Adaptive voxel filter keeps the point count and callback time consistent across environments and lidar configurations (Sec. IV-C2; Figs. 6-8)","Sliding-window maps cut memory by 38.88% to 87.76% relative to the LOCUS 1.0 octree baseline (Table II)","Open-source code and an 11 h, 16 km underground dataset with Husky and Spot data (abstract; Sec. IV-A)",[30,31,32,33,34,35],"GICP from normals raises mean and max APE by 10.82% on average (5.23% without tunnel dataset F) because normals come from sparse clouds and are not recomputed (Sec. IV-C1)","Sliding-window maps increase CPU use by 9.36% to 50.42% relative to the static octree (Table II)","ikd-tree insertion and search take on average 222% and 140% more time than the octree (Sec. IV-D1)","Smaller sliding maps bound memory but give higher APE (Sec. IV-D2)","Memory use is higher than FAST-LIO and LINS on most datasets (Table III; Sec. IV-E)","Ground truth is LOCUS 1.0 run against a survey-grade map with manual post-processing, not an independent trajectory measurement (Sec. IV-A)",[37,38,39],"one or more 3D LiDARs merged in the body frame (three Velodyne VLP16 on Husky; one lidar on Spot)","IMU for per-lidar motion distortion correction","optional non-lidar odometry (wheel-inertial, kinematic-inertial or visual-inertial) as initial guess through the sensor integration module (LiDAR-centric, loosely coupled)",[41,42],"wheeled UGV (Husky)","legged (Spot)","multi-stage GICP registration, scan-to-scan then scan-to-submap, with point covariances built from precomputed normals (GICP from normals); health-aware loosely coupled use of optional non-lidar odometry as the scan-to-scan initial guess (Sec. III; Sec. III-A)","GICP nearest-neighbour correspondences (maximum correspondence distance 0.3, 20 iterations) on points reduced by an adaptive voxel grid filter that holds the point count near a set value (1000 to 10000 tested) (Sec. III-B; Sec. IV-C)","discrete poses","IMU-based motion distortion correction of each lidar stream in the preprocessor (Sec. III)","none (odometry system; no loop closure described)","none","sliding-window robot-centred point map (window 50 m) stored in a multi-threaded octree (two threads alternately box-filter and rebuild) or an ikd-tree; normals are stored with map points (Sec. III-C; Sec. IV-D)","none for odometry; the evaluation ground truth uses survey-grade maps","6-DoF odometry and a local 3D point cloud map","Husky runs 4 threads and Spot 1 thread; in Table III LOCUS 2.0 uses 61.05 to 119.00% CPU in the column printed as max (100% = one core) and 1.01 to 2.42 GB maximum memory; GICP from normals reduced the computational metrics by 18.57% on average and raised the odometry rate by 11.10% (Sec. IV-B; Sec. IV-C1; Table III)","https:\u002F\u002Fgithub.com\u002FNeBula-Autonomy\u002FLOCUS","MIT (LICENSE file read)",[56,60,64,68],{"relation":57,"title":58,"doi_or_url":59},"preprint","LOCUS 2.0 (arXiv v2, accepted RA-L version; arXiv title adds 'Underground')","https:\u002F\u002Farxiv.org\u002Fabs\u002F2205.11784",{"relation":61,"title":62,"doi_or_url":63},"predecessor_method","LOCUS: A Multi-Sensor Lidar-Centric Solution for High-Precision Odometry and 3D Mapping in Real-Time (RA-L 2021; not read)","https:\u002F\u002Fdoi.org\u002F10.1109\u002FLRA.2020.3044864",{"relation":65,"title":66,"doi_or_url":67},"dataset","NeBula odometry dataset released with the paper","https:\u002F\u002Fgithub.com\u002FNeBula-Autonomy\u002Fnebula-odometry-dataset",{"relation":69,"title":70,"doi_or_url":53},"code_release","NeBula-Autonomy\u002FLOCUS",{"id":5,"kind":72,"shortName":7,"title":8,"authors":73,"year":9,"venue":81,"venueType":82,"publisher":83,"volumeIssuePages":84,"doi":85,"arxivId":86,"url":87,"firstPublicDate":88,"publicationStatus":16,"metadataStatus":89,"fulltextStatus":15,"era":10,"classicReason":90,"codeUrl":53,"cluster":11,"topics":91,"mdpi":92,"verification":93,"label":6,"fulltextRoute":94,"versionRead":95,"addedByCensus":96},"method",[74,75,76,77,78,79,80],"Andrzej Reinke","Matteo Palieri","Benjamin Morrell","Yun Chang","Kamak Ebadi","Luca Carlone","Ali-Akbar Agha-Mohammadi","IEEE Robotics and Automation Letters","journal","IEEE","7(4):9043-9050","10.1109\u002Flra.2022.3181357","2205.11784","https:\u002F\u002Fdoi.org\u002F10.1109\u002FLRA.2022.3181357","2022-05-24","metadata_verified","not_applicable",[11],false,"corrected","arXiv","arXiv v2 (2022-06-13), RA-L accepted preprint ('Accepted June, 2022'); IEEE version of record not read",true,[98,106,110,115,119,123,126,131,134],{"category":99,"model":100,"canonical":101,"role":102,"dataset":103,"specs":104,"locator":105},"lidar","VLP16","Velodyne VLP-16","method input","NeBula odometry dataset (DARPA SubT, Team CoSTAR)","three per Husky, extrinsically calibrated: one flat, one pitched forward 30 deg, one pitched backward 30 deg; 10 Hz (two used where marked in Table I)","Sec. IV-A; Table I",{"category":99,"model":107,"canonical":107,"role":102,"dataset":103,"specs":108,"locator":109},"Spot on-board lidar (model not reported)","one lidar, extrinsically calibrated; 10 Hz","Sec. IV-A",{"category":111,"model":112,"canonical":112,"role":102,"dataset":103,"specs":113,"locator":114},"imu","IMU (model not reported)","recorded at 50 Hz; used for motion distortion correction","Sec. III; Sec. IV-A",{"category":116,"model":117,"canonical":117,"role":102,"dataset":103,"specs":118,"locator":109},"wheel_or_leg_odometry","wheeled inertial odometry (WIO) on Husky","recorded at 50 Hz",{"category":116,"model":120,"canonical":120,"role":121,"dataset":103,"specs":122,"locator":109},"Spot kinematic inertial odometry (KIO) and visual inertial odometry (VIO), out of the box","dataset sensor","recorded in the dataset",{"category":124,"model":125,"canonical":125,"role":121,"dataset":103,"specs":122,"locator":109},"camera","camera streams (models not reported)",{"category":127,"model":128,"canonical":128,"role":102,"dataset":103,"specs":129,"locator":130},"platform","Husky","skid-steer wheeled robot on rough terrain","Sec. IV-A; Fig. 4",{"category":127,"model":132,"canonical":132,"role":102,"dataset":103,"specs":133,"locator":130},"Spot","legged robot",{"category":135,"model":136,"canonical":136,"role":137,"dataset":103,"specs":138,"locator":109},"other","survey-grade 3D map (provided by DARPA or produced by the team; instrument not reported)","reference or ground truth","ground-truth trajectory produced by LOCUS 1.0 scan-to-survey-map registration with manual post-processing",[],{"totalRows":141,"groupCount":142,"groups":143,"others":494},44,4,[144,383,430,467],{"slug":145,"group":146,"sourceId":5,"sourceLabel":6,"table":147,"selfRows":148,"metrics":149,"seqs":167,"entrants":185,"cells":193,"outcomes":375,"locators":377,"hardware":378,"wordings":380,"notes":381},"locus2-2022-table-iii","locus2_2022:Table III","Table III",30,[150,155,159,162,164],{"label":151,"unit":152,"statistic":153,"alignment":154},"APE max [m]","m","max","not_reported",{"label":156,"unit":157,"statistic":158,"alignment":154},"APE mean [%] (unit as printed)","% (as printed)","mean",{"label":160,"unit":161,"statistic":153,"alignment":90},"CPU [%] max (as printed)","% (100% = one core)",{"label":163,"unit":161,"statistic":158,"alignment":90},"CPU [%] mean (as printed)",{"label":165,"unit":166,"statistic":153,"alignment":90},"max memory [GB]","GB",[168,171,174,177,180,183],{"dataset":103,"sequence":169,"environment":170},"A: power plant, Elma WA (urban), Husky, 631.53 m","feature-poor corridors, large open spaces",{"dataset":103,"sequence":172,"environment":173},"C: power plant, Elma WA (urban), Husky, 757.40 m","feature-poor corridors, large and narrow spaces",{"dataset":103,"sequence":175,"environment":176},"F: Bruceton Mine, Pittsburgh PA (tunnel), Husky, 1569.73 m","self-similar repetitive geometry",{"dataset":103,"sequence":178,"environment":179},"H: Subway Station, Los Angeles CA (urban), Spot, 1777.45 m","3-level, multiple stairs, feature-poor corridors",{"dataset":103,"sequence":181,"environment":182},"I: Kentucky Underground Limestone Mine KY (cave), Spot, 768.82 m","large area, degraded lighting",{"dataset":103,"sequence":184,"environment":182},"J: Kentucky Underground Limestone Mine KY (cave), Husky, 2339.81 m",[186,187,190],{"name":7,"methodId":5,"linkable":96,"proposed":96,"self":96},{"name":188,"methodId":189,"linkable":96,"proposed":92,"self":92},"FAST-LIO","fastlio2021",{"name":191,"methodId":192,"linkable":96,"proposed":92,"self":92},"LINS","lins2020",[194,198,201,204,207,209,211,213,215,217,219,221,223,225,227,229,231,233,235,237,239,241,243,245,247,249,250,252,254,256,258,260,262,264,266,268,270,272,274,276,278,280,282,284,286,287,289,291,293,295,297,299,301,303,305,307,309,311,313,315,317,319,321,323,325,327,328,330,332,334,336,338,340,342,344,345,347,349,351,353,355,357,359,361,363,365,367,369,371,373],[195,195,195,196,197,195,197,197,195],0,0.19,-1,[195,199,195,200,197,195,197,197,195],1,0.09,[195,202,195,203,197,195,195,197,195],2,102.38,[195,205,195,206,197,195,195,197,195],3,185.5,[195,142,195,208,197,195,195,197,195],1.06,[199,195,195,210,197,195,197,197,195],0.79,[199,199,195,212,197,195,197,197,195],0.3,[199,202,195,214,197,195,195,197,195],89.11,[199,205,195,216,197,195,195,197,195],126.4,[199,142,195,218,197,195,195,197,195],0.36,[202,195,195,220,197,195,197,197,195],0.43,[202,199,195,222,197,195,197,197,195],0.18,[202,202,195,224,197,195,195,197,195],40.84,[202,205,195,226,197,195,195,197,195],81.5,[202,142,195,228,197,195,195,197,195],0.42,[195,195,199,230,197,195,197,197,195],0.16,[195,199,199,232,197,195,197,197,195],0.24,[195,202,199,234,197,195,195,197,195],114.79,[195,205,199,236,197,195,195,197,195],198,[195,142,199,238,197,195,195,197,195],1.3,[199,195,199,240,197,195,197,197,195],2.21,[199,199,199,242,197,195,197,197,195],4.22,[199,202,199,244,197,195,195,197,195],76.46,[199,205,199,246,197,195,195,197,195],307.2,[199,142,199,248,197,195,195,197,195],0.99,[202,195,199,220,197,195,197,197,195],[202,199,199,251,197,195,197,197,195],0.6,[202,202,199,253,197,195,195,197,195],38.43,[202,205,199,255,197,195,195,197,195],75.3,[202,142,199,257,197,195,195,197,195],0.47,[195,195,202,259,197,195,197,197,195],0.67,[195,199,202,261,197,195,197,197,195],0.45,[195,202,202,263,197,195,195,197,195],119,[195,205,202,265,197,195,195,197,195],229.2,[195,142,202,267,197,195,195,197,195],1.98,[199,195,202,269,195,195,197,197,195],48555.33,[199,199,202,271,195,195,197,197,195],9268.71,[199,202,202,273,197,195,195,197,195],156.73,[199,205,202,275,197,195,195,197,195],401.3,[199,142,202,277,197,195,195,197,195],11.31,[202,195,202,279,195,195,197,197,195],52.73,[202,199,202,281,195,195,197,197,195],23.35,[202,202,202,283,197,195,195,197,195],28.1,[202,205,202,285,197,195,195,197,195],52.3,[202,142,202,257,197,195,195,197,195],[195,195,205,288,197,195,197,197,195],0.57,[195,199,205,290,197,195,197,197,195],0.23,[195,202,205,292,197,195,195,197,195],61.05,[195,205,205,294,197,195,195,197,195],169.9,[195,142,205,296,197,195,195,197,195],2.42,[199,195,205,298,197,195,197,197,195],5.92,[199,199,205,300,197,195,197,197,195],5.69,[199,202,205,302,197,195,195,197,195],75.15,[199,205,205,304,197,195,195,197,195],160.8,[199,142,205,306,197,195,195,197,195],0.62,[202,195,205,308,197,195,197,197,195],12.11,[202,199,205,310,197,195,197,197,195],8.05,[202,202,205,312,197,195,195,197,195],39.19,[202,205,205,314,197,195,195,197,195],97.9,[202,142,205,316,197,195,195,197,195],0.61,[195,195,142,318,197,195,197,197,195],1.39,[195,199,142,320,197,195,197,197,195],1.95,[195,202,142,322,197,195,195,197,195],72.11,[195,205,142,324,197,195,195,197,195],141.6,[195,142,142,326,197,195,195,197,195],1.01,[199,195,142,248,197,195,197,197,195],[199,199,142,329,197,195,197,197,195],1.44,[199,202,142,331,197,195,195,197,195],117.87,[199,205,142,333,197,195,195,197,195],167.8,[199,142,142,335,197,195,195,197,195],0.8,[202,195,142,337,197,195,197,197,195],0.86,[202,199,142,339,197,195,197,197,195],0.85,[202,202,142,341,197,195,195,197,195],75.9,[202,205,142,343,197,195,195,197,195],101.4,[202,142,142,339,197,195,195,197,195],[195,195,346,296,197,195,197,197,195],5,[195,199,346,348,197,195,197,197,195],3.88,[195,202,346,350,197,195,195,197,195],107.72,[195,205,346,352,197,195,195,197,195],185,[195,142,346,354,197,195,195,197,195],2.13,[199,195,346,356,197,195,197,197,195],1.72,[199,199,346,358,197,195,197,197,195],2.6,[199,202,346,360,197,195,195,197,195],126.72,[199,205,346,362,197,195,195,197,195],332.5,[199,142,346,364,197,195,195,197,195],2.54,[202,195,346,366,197,195,197,197,195],3.56,[202,199,346,368,197,195,197,197,195],5.79,[202,202,346,370,197,195,195,197,195],73.76,[202,205,346,372,197,195,195,197,195],176.5,[202,142,346,374,197,195,195,197,195],1.85,[376],"failed (authors state only LOCUS 2.0 does not fail in tunnel dataset F)",[147],[379],"not_reported (the computer used for Table III is not stated in the paper)",[],[382],"Underground datasets A, C, F, H, I, J (Table I); LOCUS 2.0 versus FAST-LIO and LINS; column labels reproduced as printed (APE max [m], APE mean [%], CPU [%] max and mean, max memory [GB]); many printed 'max' values are below 'mean' values; ground truth from LOCUS 1.0 against survey-grade maps",{"slug":384,"group":385,"sourceId":5,"sourceLabel":6,"table":386,"selfRows":387,"metrics":388,"seqs":394,"entrants":398,"cells":407,"outcomes":424,"locators":425,"hardware":426,"wordings":427,"notes":428},"locus2-2022-table-ii","locus2_2022:Table II","Table II",8,[389,392],{"label":390,"unit":391,"statistic":154,"alignment":90},"Memory (relative change)","% change",{"label":393,"unit":391,"statistic":154,"alignment":90},"CPU (relative change)",[395],{"dataset":103,"sequence":396,"environment":397},"datasets used in Sec. IV-D (F and I shown in Fig. 9)","underground",[399,401,403,405],{"name":400,"methodId":5,"linkable":96,"proposed":96,"self":96},"LOCUS 2.0 with ikd-tree",{"name":402,"methodId":5,"linkable":96,"proposed":96,"self":96},"LOCUS 2.0 with mto 0.001 (multi-threaded octree, leaf 0.001 m)",{"name":404,"methodId":5,"linkable":96,"proposed":96,"self":96},"LOCUS 2.0 with mto 0.01",{"name":406,"methodId":5,"linkable":96,"proposed":96,"self":96},"LOCUS 2.0 with mto 0.1",[408,410,412,414,416,418,420,422],[195,195,195,409,197,195,197,197,195],-68.09,[195,199,195,411,197,195,197,197,195],9.36,[199,195,195,413,197,195,197,197,195],-38.88,[199,199,195,415,197,195,197,197,195],50.42,[202,195,195,417,197,195,197,197,195],-62.15,[202,199,195,419,197,195,197,197,195],44.36,[205,195,195,421,197,195,197,197,195],-87.76,[205,199,195,423,197,195,197,197,195],19.61,[],[386],[],[],[429],"Relative memory and CPU change of sliding-window map structures versus the LOCUS 1.0 static octree with 0.001 m leaf (baseline), 50 m map window, GICP from normals",{"slug":431,"group":432,"sourceId":5,"sourceLabel":6,"table":433,"selfRows":142,"metrics":434,"seqs":443,"entrants":448,"cells":451,"outcomes":460,"locators":461,"hardware":463,"wordings":464,"notes":465},"locus2-2022-text-sec-iv-c1","locus2_2022:Text Sec. IV-C1","Text Sec. IV-C1",[435,437,439,441],{"label":436,"unit":391,"statistic":158,"alignment":90},"average reduction of computational metrics (CPU, delay, registration and callback times)",{"label":438,"unit":391,"statistic":158,"alignment":90},"odometry update rate increase",{"label":440,"unit":391,"statistic":158,"alignment":154},"mean and max APE increase",{"label":442,"unit":391,"statistic":158,"alignment":154},"APE increase without dataset F",[444,446],{"dataset":103,"sequence":445,"environment":397},"A-J (average)",{"dataset":103,"sequence":447,"environment":397},"A-J excluding tunnel F (average)",[449],{"name":450,"methodId":5,"linkable":96,"proposed":96,"self":96},"LOCUS 2.0, GICP from normals vs GICP",[452,454,456,458],[195,195,195,453,197,195,197,197,195],-18.57,[195,199,195,455,197,195,197,197,195],11.1,[195,202,195,457,197,195,197,197,195],10.82,[195,205,199,459,197,195,197,197,195],5.23,[],[462],"Sec. IV-C1",[],[],[466],"GICP from normals versus standard GICP inside LOCUS 2.0, averaged over datasets A-J and 5 runs each; percentage changes stated in text",{"slug":468,"group":469,"sourceId":5,"sourceLabel":6,"table":470,"selfRows":202,"metrics":471,"seqs":477,"entrants":480,"cells":482,"outcomes":487,"locators":488,"hardware":490,"wordings":491,"notes":492},"locus2-2022-text-sec-iv-d1","locus2_2022:Text Sec. IV-D1","Text Sec. IV-D1",[472,475],{"label":473,"unit":474,"statistic":158,"alignment":90},"insertion computation time relative to octree","% more",{"label":476,"unit":474,"statistic":158,"alignment":90},"search computation time relative to octree",[478],{"dataset":103,"sequence":479,"environment":397},"all datasets (average)",[481],{"name":400,"methodId":5,"linkable":96,"proposed":96,"self":96},[483,485],[195,195,195,484,197,195,197,197,195],222,[195,199,195,486,197,195,197,197,195],140,[],[489],"Sec. IV-D1",[],[],[493],"Map structure operation timing relative to the octree",[],1790510661684]