[{"data":1,"prerenderedAt":550},["ShallowReactive",2],{"method-lioekf2024":3},{"method":4,"reference":58,"equipment":83,"figures":116,"results":117},{"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":39,"association":40,"timeModel":41,"deskew":42,"loopClosure":43,"globalOptimization":43,"mapRepresentation":44,"prior":45,"outputGeometry":46,"compute":47,"codeUrl":48,"codeLicense":49,"relatedVersions":50},"lioekf2024","Wu et al., 2024a","LIO-EKF","LIO-EKF: High Frequency LiDAR-Inertial Odometry using Extended Kalman Filters",2024,"recent","C05","odometry_with_local_mapping","LIO-EKF 把 KISS-ICP 的點對點配準與傳統誤差狀態擴展卡爾曼濾波結合成緊耦合 LiDAR 慣性里程計。預測步使用慣性導航領域的精確捷聯 INS 機械編排，作者認為 IMU 預測夠準，因此每個掃描只做一次卡爾曼修正，不需迭代。資料關聯的最大對應距離不以經驗設定，而是由 IMU 相對位姿不確定度（經無跡轉換投影到點距）、體素地圖離散化誤差與 LiDAR 測距雜訊三者組合，以三倍標準差自動求得，藉此減少需調整的參數。","Tightly coupled LIO that combines KISS-ICP point-to-point registration against a voxel map with a single-update error-state EKF driven by strapdown INS mechanization, and sets the correspondence threshold automatically from IMU relative-pose uncertainty, voxel discretization error and range noise.","full_text_reviewed","peer_reviewed_published","supplementary","原論文只在都市車載、校園地面機器人與手持校園資料上評估，未涉及營建場域。在 SubT-MRS（ICCV 2023 SLAM 挑戰 LiDAR 組）中，一個以 LIO-EKF 為基礎的參賽方法在長走廊與多樓層序列的 ATE 分別為 2.99 m 與 5.5 m，隧道序列為 0.22 m [zhao2024subtmrs]；這些室內退化場景與施工中建物走廊、樓梯相近，顯示僅靠點對點配準加單次 EKF 更新在幾何退化處仍可能產生公尺級誤差（推論）。",[20,21],"public_benchmark","cross_site",[23,24,25,26],"Accuracy on par with FAST-LIO2 and LIO-SAM on UrbanNav, M2DGR and Newer College with one configuration; better relative errors than FAST-LIO2 on UrbanNav and Newer College (Table I)","About two times faster than FAST-LIO2 and four times faster than LIO-SAM in scan processing, close to a 100 Hz IMU rate (Table II)","Iterating the update 10 or 100 times changed relative translation error by at most 0.08 percentage points, while processing time grew 5.6 to 6.8 times at 100 iterations in Table III (the text states a factor of 8) (Table III, Sec. IV-E-1)","Adaptive threshold generalized better than fixed 0.3 m or 1 m thresholds and the KISS-ICP adaptive threshold across UrbanNav and M2DGR (Table IV)",[28,29,30,31],"Absolute translation error on UrbanNav was higher than LIO-SAM on all three sequences and higher than FAST-LIO2 on 20210517 (24.73 vs 17.62 m); ATE rotation on 20210518 was 7.46 deg versus about 3 deg for both baselines (Table I)","Tunable parameters remain: maximum points per voxel, voxel size, initial state covariance and point observation covariance (Sec. IV-B)","Odometry only, without loop closure (Sec. III)","LIO-SAM could not be run on Newer College because it needs IMU attitude output (Sec. IV-C)",[33,34],"3D LiDAR (dataset sensors; models not named in the paper)","consumer-grade MEMS IMU",[36,37,38],"vehicle (UrbanNav, Hong Kong)","wheeled UGV (M2DGR, Shanghai campus)","handheld (Newer College, Oxford)","classical error-state EKF with a single correction per scan (no iterations), strapdown INS mechanization for prediction and the Kalman-gain reformulation of FAST-LIO for the update (Sec. III-A, III-B)","KISS-ICP-style point-to-point: voxel-subsampled scan matched to a voxel-hashed local map by nearest neighbour within an adaptive threshold tau = 3 sqrt(sigma_p2p^2 + sigma_map^2 + sigma_range^2) that combines IMU relative-pose uncertainty (unscented transform), voxel discretization error and LiDAR range noise (Sec. III-B, III-C)","discrete scan poses with IMU prediction between scans (Sec. III-A)","each scan deskewed with the IMU-predicted pose before registration (Sec. III-B)","none","voxel-grid local point map with a maximum number of points per voxel, maintained as in KISS-ICP (Sec. III-B)","LiDAR-IMU extrinsic calibration; IMU and range noise taken from datasheets (Sec. III-B, IV-B)","odometry at close to IMU rate and a local voxel point map","CPU: average correction-step time 12.37, 13.65 and 7.36 ms on UrbanNav, M2DGR and Newer College versus 24.03, 29.96 and 9.42 ms for FAST-LIO2 on an Intel i7-10700 at 2.90 GHz with 32 GB RAM (Table II)","https:\u002F\u002Fgithub.com\u002FYibinWu\u002FLIO-EKF","MIT (LICENSE file checked)",[51,55],{"relation":52,"title":53,"doi_or_url":54},"preprint","arXiv 2311.09887 (v1 2023-11-16, v2 2024-05-08)","https:\u002F\u002Farxiv.org\u002Fabs\u002F2311.09887",{"relation":56,"title":57,"doi_or_url":48},"code_release","YibinWu\u002FLIO-EKF",{"id":5,"kind":59,"shortName":7,"title":8,"authors":60,"year":9,"venue":67,"venueType":68,"publisher":69,"volumeIssuePages":70,"doi":71,"arxivId":72,"url":73,"firstPublicDate":74,"publicationStatus":16,"metadataStatus":75,"fulltextStatus":15,"era":10,"classicReason":76,"codeUrl":48,"cluster":11,"topics":77,"mdpi":78,"verification":79,"label":6,"fulltextRoute":80,"versionRead":81,"addedByCensus":82},"method",[61,62,63,64,65,66],"Yibin Wu","Tiziano Guadagnino","Louis Wiesmann","Lasse Klingbeil","Cyrill Stachniss","Heiner Kuhlmann","2024 IEEE International Conference on Robotics and Automation (ICRA)","conference","IEEE","pp. 13741-13747","10.1109\u002Ficra57147.2024.10610667","2311.09887","https:\u002F\u002Fdoi.org\u002F10.1109\u002FICRA57147.2024.10610667","2023-11-16","metadata_verified","not_applicable",[11],false,"corrected","arXiv","arXiv v2 (2024-05-08); IEEE ICRA 2024 version of record not read",true,[84,91,95,99,105,112],{"category":85,"model":86,"canonical":86,"role":87,"dataset":88,"specs":89,"locator":90},"platform","car","dataset sensor","UrbanNav","autonomous driving sequences in Hong Kong","Sec. IV-A; Fig. 1",{"category":85,"model":92,"canonical":92,"role":87,"dataset":93,"specs":94,"locator":90},"wheeled mobile robot","M2DGR","university campus in Shanghai",{"category":85,"model":96,"canonical":96,"role":87,"dataset":97,"specs":98,"locator":90},"handheld device","Newer College Dataset","Oxford university campus",{"category":100,"model":101,"canonical":101,"role":87,"dataset":102,"specs":103,"locator":104},"imu","consumer-grade MEMS IMUs (models not stated)","UrbanNav, M2DGR and Newer College","not_reported","Sec. IV-A",{"category":106,"model":107,"canonical":107,"role":108,"dataset":109,"specs":110,"locator":111},"compute","Intel i7-10700","compute for runtime",null,"2.90 GHz, 32 GB RAM, desktop","Sec. IV-D",{"category":113,"model":114,"canonical":114,"role":87,"dataset":102,"specs":103,"locator":115},"lidar","3D LiDAR (dataset sensors; models not stated)","Sec. III; Sec. IV-A",[],{"totalRows":118,"groupCount":119,"groups":120,"others":549},85,4,[121,363,446,509],{"slug":122,"group":123,"sourceId":5,"sourceLabel":6,"table":124,"selfRows":125,"metrics":126,"seqs":140,"entrants":160,"cells":168,"outcomes":356,"locators":358,"hardware":359,"wordings":360,"notes":361},"lioekf2024-table-i","lioekf2024:Table I","Table I",32,[127,131,134,137],{"label":128,"unit":129,"statistic":130,"alignment":103},"Avg. tra. (KITTI relative translation error)","%","mean",{"label":132,"unit":133,"statistic":130,"alignment":103},"Avg. rot. (KITTI relative rotation error)","deg\u002Fm (as printed)",{"label":135,"unit":136,"statistic":103,"alignment":103},"ATE. tra. (unit taken as m from the column name; the Table I footnote lists the two ATE units in swapped order)","m",{"label":138,"unit":139,"statistic":103,"alignment":103},"ATE. rot. (unit taken as deg from the column name; the Table I footnote lists the two ATE units in swapped order)","deg",[141,144,146,148,151,153,155,158],{"dataset":88,"sequence":142,"environment":143},"20210517","urban driving, Hong Kong",{"dataset":88,"sequence":145,"environment":143},"20210518",{"dataset":88,"sequence":147,"environment":143},"20210521",{"dataset":93,"sequence":149,"environment":150},"street 01-05 (average)","campus streets, wheeled robot",{"dataset":93,"sequence":152,"environment":150},"street 06",{"dataset":93,"sequence":154,"environment":150},"street 08",{"dataset":97,"sequence":156,"environment":157},"short exp","handheld, Oxford college",{"dataset":97,"sequence":159,"environment":157},"long exp",[161,164,167],{"name":162,"methodId":163,"linkable":82,"proposed":78,"self":78},"FAST-LIO2","fastlio2_2022",{"name":165,"methodId":166,"linkable":82,"proposed":78,"self":78},"LIO-SAM","liosam2020",{"name":7,"methodId":5,"linkable":82,"proposed":82,"self":82},[169,173,176,179,182,184,186,188,190,192,194,196,198,200,202,204,206,208,210,212,214,216,218,220,222,224,226,228,230,232,234,236,238,240,242,244,246,247,249,250,252,254,256,258,260,261,263,265,267,269,271,273,275,277,279,281,283,285,287,289,291,294,295,297,299,301,303,305,307,309,311,313,315,318,320,322,324,325,326,327,328,330,332,334,336,339,341,343,345,346,347,348,349,351,353,354],[170,170,170,171,172,170,172,172,170],0,4.11,-1,[170,174,170,175,172,170,172,172,170],1,1.68,[170,177,170,178,172,170,172,172,170],2,17.62,[170,180,170,181,172,170,172,172,170],3,4.45,[174,170,170,183,172,170,172,172,170],3.18,[174,174,170,185,172,170,172,172,170],1.4,[174,177,170,187,172,170,172,172,170],20.74,[174,180,170,189,172,170,172,172,170],4.2,[177,170,170,191,172,170,172,172,170],3.2,[177,174,170,193,172,170,172,172,170],1.45,[177,177,170,195,172,170,172,172,170],24.73,[177,180,170,197,172,170,172,172,170],5.05,[170,170,174,199,172,170,172,172,170],2.73,[170,174,174,201,172,170,172,172,170],1.3,[170,177,174,203,172,170,172,172,170],23.02,[170,180,174,205,172,170,172,172,170],3.27,[174,170,174,207,172,170,172,172,170],2.52,[174,174,174,209,172,170,172,172,170],1.31,[174,177,174,211,172,170,172,172,170],20.37,[174,180,174,213,172,170,172,172,170],2.98,[177,170,174,215,172,170,172,172,170],2.2,[177,174,174,217,172,170,172,172,170],1.14,[177,177,174,219,172,170,172,172,170],22.44,[177,180,174,221,172,170,172,172,170],7.46,[170,170,177,223,172,170,172,172,170],3.56,[170,174,177,225,172,170,172,172,170],1.63,[170,177,177,227,172,170,172,172,170],47.29,[170,180,177,229,172,170,172,172,170],5.18,[174,170,177,231,172,170,172,172,170],2.94,[174,174,177,233,172,170,172,172,170],1.62,[174,177,177,235,172,170,172,172,170],30.98,[174,180,177,237,172,170,172,172,170],4.51,[177,170,177,239,172,170,172,172,170],2.96,[177,174,177,241,172,170,172,172,170],1.54,[177,177,177,243,172,170,172,172,170],34.97,[177,180,177,245,172,170,172,172,170],4.47,[170,170,180,233,172,170,172,172,170],[170,174,180,248,172,170,172,172,170],0.79,[170,177,180,197,172,170,172,172,170],[170,180,180,251,172,170,172,172,170],1.79,[174,170,180,253,172,170,172,172,170],3.15,[174,174,180,255,172,170,172,172,170],1.52,[174,177,180,257,172,170,172,172,170],10.19,[174,180,180,259,172,170,172,172,170],4.27,[177,170,180,175,172,170,172,172,170],[177,174,180,262,172,170,172,172,170],0.83,[177,177,180,264,172,170,172,172,170],5.33,[177,180,180,266,172,170,172,172,170],1.7,[170,170,119,268,172,170,172,172,170],3.41,[170,174,119,270,172,170,172,172,170],1.55,[170,177,119,272,172,170,172,172,170],8.93,[170,180,119,274,172,170,172,172,170],2.29,[174,170,119,276,172,170,172,172,170],3.65,[174,174,119,278,172,170,172,172,170],1.65,[174,177,119,280,172,170,172,172,170],9.04,[174,180,119,282,172,170,172,172,170],2.41,[177,170,119,284,172,170,172,172,170],3.37,[177,174,119,286,172,170,172,172,170],1.56,[177,177,119,288,172,170,172,172,170],9.05,[177,180,119,290,172,170,172,172,170],2.27,[170,170,292,293,172,170,172,172,170],5,1.1,[170,174,292,278,172,170,172,172,170],[170,177,292,296,172,170,172,172,170],2.12,[170,180,292,298,172,170,172,172,170],1.71,[174,170,292,300,172,170,172,172,170],3.73,[174,174,292,302,172,170,172,172,170],5.78,[174,177,292,304,172,170,172,172,170],4.21,[174,180,292,306,172,170,172,172,170],6.36,[177,170,292,308,172,170,172,172,170],1.28,[177,174,292,310,172,170,172,172,170],1.85,[177,177,292,312,172,170,172,172,170],2.22,[177,180,292,314,172,170,172,172,170],1.88,[170,170,316,317,172,170,172,172,170],6,1.05,[170,174,316,319,172,170,172,172,170],1.01,[170,177,316,321,172,170,172,172,170],5.14,[170,180,316,323,172,170,172,172,170],2.49,[174,170,316,109,170,170,172,172,170],[174,174,316,109,170,170,172,172,170],[174,177,316,109,170,170,172,172,170],[174,180,316,109,170,170,172,172,170],[177,170,316,329,172,170,172,172,170],0.63,[177,174,316,331,172,170,172,172,170],0.73,[177,177,316,333,172,170,172,172,170],4.16,[177,180,316,335,172,170,172,172,170],1.75,[170,170,337,338,172,170,172,172,170],7,1.09,[170,174,337,340,172,170,172,172,170],1.33,[170,177,337,342,172,170,172,172,170],6.33,[170,180,337,344,172,170,172,172,170],4.22,[174,170,337,109,170,170,172,172,170],[174,174,337,109,170,170,172,172,170],[174,177,337,109,170,170,172,172,170],[174,180,337,109,170,170,172,172,170],[177,170,337,350,172,170,172,172,170],0.74,[177,174,337,352,172,170,172,172,170],0.91,[177,177,337,321,172,170,172,172,170],[177,180,337,355,172,170,172,172,170],2.34,[357],"not_run",[124],[],[],[362],"Default parameters for FAST-LIO2 and LIO-SAM, LIO-SAM loop closure disabled; one LIO-EKF configuration for all data; KITTI relative errors and ATE; LIO-SAM not run on Newer College (needs IMU attitude)",{"slug":364,"group":365,"sourceId":5,"sourceLabel":6,"table":366,"selfRows":125,"metrics":367,"seqs":372,"entrants":378,"cells":387,"outcomes":440,"locators":441,"hardware":442,"wordings":443,"notes":444},"lioekf2024-table-iv","lioekf2024:Table IV","Table IV",[368,370],{"label":369,"unit":129,"statistic":130,"alignment":103},"Avg. tra.",{"label":371,"unit":133,"statistic":130,"alignment":103},"Avg. rot.",[373,374,375,377],{"dataset":88,"sequence":145,"environment":143},{"dataset":88,"sequence":147,"environment":143},{"dataset":93,"sequence":376,"environment":150},"street 01",{"dataset":93,"sequence":154,"environment":150},[379,381,383,385],{"name":380,"methodId":5,"linkable":82,"proposed":78,"self":82},"LIO-EKF with fixed threshold 0.3 m",{"name":382,"methodId":5,"linkable":82,"proposed":78,"self":82},"LIO-EKF with fixed threshold 1 m",{"name":384,"methodId":5,"linkable":82,"proposed":78,"self":82},"LIO-EKF with KISS-ICP adaptive threshold",{"name":386,"methodId":5,"linkable":82,"proposed":82,"self":82},"LIO-EKF with proposed adaptive threshold",[388,390,392,394,396,398,399,400,401,402,404,406,407,408,410,411,412,414,416,418,420,421,423,425,427,429,431,432,434,436,438,439],[170,170,170,389,172,170,172,172,170],2.23,[170,174,170,391,172,170,172,172,170],1.25,[174,170,170,393,172,170,172,172,170],2.18,[174,174,170,395,172,170,172,172,170],1.12,[177,170,170,397,172,170,172,172,170],2.24,[177,174,170,217,172,170,172,172,170],[180,170,170,215,172,170,172,172,170],[180,174,170,217,172,170,172,172,170],[170,170,174,239,172,170,172,172,170],[170,174,174,403,172,170,172,172,170],1.72,[174,170,174,405,172,170,172,172,170],2.95,[174,174,174,270,172,170,172,172,170],[177,170,174,405,172,170,172,172,170],[177,174,174,409,172,170,172,172,170],1.59,[180,170,174,239,172,170,172,172,170],[180,174,174,241,172,170,172,172,170],[170,170,177,413,172,170,172,172,170],4.25,[170,174,177,415,172,170,172,172,170],2.26,[174,170,177,417,172,170,172,172,170],1.46,[174,174,177,419,172,170,172,172,170],0.86,[177,170,177,251,172,170,172,172,170],[177,174,177,422,172,170,172,172,170],1.04,[180,170,177,424,172,170,172,172,170],1.44,[180,174,177,426,172,170,172,172,170],0.85,[170,170,180,428,172,170,172,172,170],2.66,[170,174,180,430,172,170,172,172,170],2.58,[174,170,180,340,172,170,172,172,170],[174,174,180,433,172,170,172,172,170],1.93,[177,170,180,435,172,170,172,172,170],2.06,[177,174,180,437,172,170,172,172,170],2.5,[180,170,180,308,172,170,172,172,170],[180,174,180,310,172,170,172,172,170],[],[366],[],[],[445],"Ablation: fixed correspondence thresholds (0.3 m, 1 m), KISS-ICP adaptive threshold, and the proposed threshold; KITTI relative errors",{"slug":447,"group":448,"sourceId":5,"sourceLabel":6,"table":449,"selfRows":450,"metrics":451,"seqs":457,"entrants":461,"cells":468,"outcomes":502,"locators":503,"hardware":504,"wordings":506,"notes":507},"lioekf2024-table-iii","lioekf2024:Table III","Table III",18,[452,453,454],{"label":369,"unit":129,"statistic":130,"alignment":103},{"label":371,"unit":133,"statistic":130,"alignment":103},{"label":455,"unit":456,"statistic":130,"alignment":103},"Processing Time","ms",[458,460],{"dataset":93,"sequence":459,"environment":150},"street 05",{"dataset":97,"sequence":156,"environment":157},[462,464,466],{"name":463,"methodId":5,"linkable":82,"proposed":82,"self":82},"LIO-EKF (1 iteration)",{"name":465,"methodId":5,"linkable":82,"proposed":78,"self":82},"LIO-EKF (10 iterations)",{"name":467,"methodId":5,"linkable":82,"proposed":78,"self":82},"LIO-EKF (100 iterations)",[469,471,473,475,477,479,481,482,484,486,487,488,490,492,494,496,498,500],[170,170,170,470,172,170,170,172,170],2.64,[170,174,170,472,172,170,170,172,170],0.8,[170,177,170,474,172,170,170,172,170],11.37,[174,170,170,476,172,170,170,172,170],2.61,[174,174,170,478,172,170,170,172,170],0.77,[174,177,170,480,172,170,170,172,170],35.1,[177,170,170,476,172,170,170,172,170],[177,174,170,483,172,170,170,172,170],0.76,[177,177,170,485,172,170,170,172,170],63.9,[170,170,174,329,172,170,170,172,170],[170,174,174,331,172,170,170,172,170],[170,177,174,489,172,170,170,172,170],12.41,[174,170,174,491,172,170,170,172,170],0.6,[174,174,174,493,172,170,170,172,170],0.62,[174,177,174,495,172,170,170,172,170],26.92,[177,170,174,497,172,170,170,172,170],0.55,[177,174,174,499,172,170,170,172,170],0.58,[177,177,174,501,172,170,170,172,170],84.21,[],[449],[505],"Intel i7-10700 2.90 GHz, 32 GB RAM",[],[508],"Ablation: LIO-EKF update with 1 (EKF), 10 or 100 iterations (IEKF); KITTI relative errors and processing time",{"slug":510,"group":511,"sourceId":5,"sourceLabel":6,"table":512,"selfRows":180,"metrics":513,"seqs":516,"entrants":521,"cells":525,"outcomes":543,"locators":544,"hardware":545,"wordings":546,"notes":547},"lioekf2024-table-ii","lioekf2024:Table II","Table II",[514],{"label":515,"unit":456,"statistic":130,"alignment":76},"average processing time",[517,519,520],{"dataset":88,"sequence":518,"environment":143},"all sequences (average)",{"dataset":93,"sequence":518,"environment":150},{"dataset":97,"sequence":518,"environment":157},[522,523,524],{"name":165,"methodId":166,"linkable":82,"proposed":78,"self":78},{"name":162,"methodId":163,"linkable":82,"proposed":78,"self":78},{"name":7,"methodId":5,"linkable":82,"proposed":82,"self":82},[526,528,530,531,533,535,537,539,541],[170,170,170,527,172,170,170,172,170],41.68,[170,170,174,529,172,170,170,172,170],60.01,[170,170,177,109,170,170,170,172,170],[174,170,170,532,172,170,170,172,170],24.03,[174,170,174,534,172,170,170,172,170],29.96,[174,170,177,536,172,170,170,172,170],9.42,[177,170,170,538,172,170,170,172,170],12.37,[177,170,174,540,172,170,170,172,170],13.65,[177,170,177,542,172,170,170,172,170],7.36,[357],[512],[505],[],[548],"Average processing time of the scan correction step, averaged over sequences of each dataset",[],1790510658601]