[{"data":1,"prerenderedAt":455},["ShallowReactive",2],{"method-in2laama2021":3},{"method":4,"reference":70,"equipment":92,"figures":127,"results":128},{"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":35,"platform":38,"estimator":41,"association":42,"timeModel":43,"deskew":44,"loopClosure":45,"globalOptimization":46,"mapRepresentation":47,"prior":48,"outputGeometry":49,"compute":50,"codeUrl":51,"codeLicense":52,"relatedVersions":53},"in2laama2021","Le Gentil et al., 2021","IN2LAAMA","IN2LAAMA: Inertial Lidar Localization Autocalibration and Mapping",2021,"recent","C05","full_slam_with_global_correction","IN2LAAMA 是以 3D LiDAR 與 6 自由度 IMU 進行離線批次定位、建圖與外參自動校正的框架。它對每個 IMU 軸以高斯過程回歸建立連續慣性訊號，再對每個 LiDAR 點的時間戳記做預積分，得到「上取樣預積分量測」（UPM），因此不需假設等速等運動模型即可精確描述掃描期間的運動並去畸變。後端把點到線、點到平面距離、影格間 IMU 因子、偏差與時間偏移因子放入同一個全批次最佳化，前端則依最新狀態重算特徵，並以大於 360 度的掃描與雙向關聯維持掃描一致性；最後可把 LiDAR 與 IMU 外參加入狀態，不需校正標靶。","Offline full-batch LiDAR-inertial localisation, mapping and targetless extrinsic and time-shift calibration that uses Gaussian-process-upsampled IMU preintegration at every point timestamp (UPMs) to remove motion distortion without a motion model, jointly optimizing point-to-line, point-to-plane, IMU, bias and time-shift factors with periodic feature recomputation and proximity loop closures.","full_text_reviewed","peer_reviewed_published","supplementary","論文未在營建工地驗證，真實資料為 UTS 實驗室、跨樓層樓梯間與校園道路。其實驗以人工分割的牆面與樓板計算點到平面 RMS 距離，並把實驗室地圖疊合於數位化的施工前平面圖，作者也指出平面圖因結構變更與家具而與現況不符；這種以平面殘差評估點雲品質的方式與營建驗收的平整度檢查概念相近（推論）。離線全批次、可自動校正外參與時間偏移的特性，適合事後處理的點雲測繪服務（作者以 3D mapping as a service 為例，Sec. VIII），但計算成本高。同作者的元件研究見 [legentil2018lidarimucalib] 與 [legentil2020gpm]。",[20,21,22],"simulation","completed_building","cross_site",[24,25,26,27,28],"All UPM variants were at least an order of magnitude more accurate than standard preintegration at the LiDAR point times, with GP regression best (Sec. VII-A, Fig. 9)","With IMU factors it handled fast simulated motion (average 125 deg\u002Fs) without failures where the authors' earlier IN2LAMA failed in 37 of 50 runs (Table I)","Staircase wall RMS point-to-plane distance 10 mm versus 31 mm for IN2LAMA and 60 mm for A-LOAM (Table VI)","Removing the constant-velocity assumption cut RMSE position error from 2.67 m to 0.087 m in fast simulated motion (Table IV)","Targetless extrinsic calibration reached about 1 cm and below 0.13 deg error in simulation, and on real data gave crisper maps (27 mm mean plane RMS) than a camera-chained calibration (58 mm) (Table V, Sec. VII-F)",[30,31,32,33,34],"Offline only: execution times are 23 to 55 times the sequence duration with several GiB of memory (Table VII, Sec. VIII)","Front-end designed for structured geometry; outdoor scenes are challenging for the feature extraction (Sec. VII-E-2)","Loop closure uses simple pose proximity and does not handle large drift or kidnapped cases (Sec. V-D)","Accelerometer biases are unobservable on translation-only segments; an extra zero-bias factor is needed (Sec. VI-D, VII-E-2)","Calibration requires trajectories that excite all six IMU axes and observe at least three non-coplanar planes or non-collinear edges (Sec. VII-F)",[36,37],"3D spinning LiDAR with per-point timestamps (Velodyne VLP-16; Velodyne HDL-32 in the MC2SLAM campus drive sequence)","6-DoF IMU (Xsens MTi-3 at 100 Hz; HDL-32 built-in IMU)",[39,40,20],"authors' self-contained VLP-16 and Xsens MTi-3 sensor suite, carrier not stated (indoor lab, staircase and calibration runs)","vehicle (car, MC2SLAM campus drive)","offline full-batch MAP optimization in Ceres with analytic Jacobians over all frame poses, velocities, per-frame IMU bias and time-shift corrections (and optionally LiDAR-IMU extrinsics), built incrementally with feature recomputation; bisquare weights on lidar residuals and Cauchy loss on lidar and IMU factors (Sec. III-B, VI)","channel-wise features from a linear-regression curvature score (cosine between left and right fitted lines), planar points and inward or outward edges binned per line; point-to-plane (3 nearest) and point-to-line (2 nearest) associations between frames with spatial-spread and patch-consistency outlier tests; frames longer than 360 deg (520 deg) with back-and-forth association (Sec. V-A, V-C)","discrete states at the start of each LiDAR frame; within frames, Upsampled Preintegrated Measurements (UPMs) from per-axis Gaussian-process regression of IMU readings give the pose at every point timestamp without an explicit motion model (Sec. III-C)","every point is projected with its own UPM inside the residuals, so motion distortion is re-corrected whenever the state changes; features are recomputed from the current estimate (Sec. IV-A, V-B, VI-A)","yes; proximity-based candidates using radial distance, height offset and z-axis angle between LiDAR frames, optional ICP fitness test, added as lidar factors (Sec. V-D)","the whole trajectory, biases, time-shifts and optionally extrinsics are estimated in one batch problem including loop-closure factors (Sec. III-B, VI-A)","dense point cloud reconstructed by projecting all points with the final trajectory and UPMs; no persistent map structure (Sec. VI-E)","no trajectory prior; Gaussian prior on inter-sensor time-shift; temporary zero prior on the initial accelerometer bias to handle translation-only starts (Sec. III-B, VI-D)","full 6-DoF trajectory, IMU biases, time-shift, LiDAR-IMU extrinsic calibration and dense motion-corrected point cloud map (Figs. 1, 11-13)","offline, not real time: 950 s for a 41 s lab sequence, 1306 s for a 35 s staircase sequence and 4699 s for an 85 s outdoor HDL-32 sequence, with 4.2 to 6.45 GiB data memory plus up to 5.77 GiB optimizer memory; UPMs need more than 340 MB per second of VLP-16 data; hardware not reported (Table VII, Sec. VI-E)",null,"not_applicable (only datasets released)",[54,58,62,66],{"relation":55,"title":56,"doi_or_url":57},"preprint","arXiv 1905.09517 (v1 2019-05-23, v2 2020-04-06, v3 2020-10-22, accepted version)","https:\u002F\u002Farxiv.org\u002Fabs\u002F1905.09517",{"relation":59,"title":60,"doi_or_url":61},"accepted_manuscript","UTS OPUS repository copy (submitted version per OpenAlex; not opened)","http:\u002F\u002Fhdl.handle.net\u002F10453\u002F147466",{"relation":63,"title":64,"doi_or_url":65},"conference_version","IN2LAMA: INertial Lidar Localisation And MApping, ICRA 2019 (predecessor without IMU factors and calibration; DOI checked in Crossref)","https:\u002F\u002Fdoi.org\u002F10.1109\u002FICRA.2019.8794429",{"relation":67,"title":68,"doi_or_url":69},"dataset","UTS-CAS\u002Fin2laama_datasets (real-data sequences with per-point timestamps)","https:\u002F\u002Fgithub.com\u002FUTS-CAS\u002Fin2laama_datasets",{"id":5,"kind":71,"shortName":7,"title":8,"authors":72,"year":9,"venue":76,"venueType":77,"publisher":78,"volumeIssuePages":79,"doi":80,"arxivId":81,"url":82,"firstPublicDate":83,"publicationStatus":16,"metadataStatus":84,"fulltextStatus":15,"era":10,"classicReason":85,"codeUrl":51,"cluster":11,"topics":86,"mdpi":87,"verification":88,"label":6,"fulltextRoute":89,"versionRead":90,"addedByCensus":91},"method",[73,74,75],"Cedric Le Gentil","Teresa Vidal-Calleja","Shoudong Huang","IEEE Transactions on Robotics","journal","IEEE","37(1):275-290","10.1109\u002Ftro.2020.3018641","1905.09517","https:\u002F\u002Fdoi.org\u002F10.1109\u002FTRO.2020.3018641","2019-05-23","metadata_verified","not_applicable",[11],false,"corrected","arXiv","arXiv v3 (2020-10-22), the accepted T-RO version with IEEE copyright notice; IEEE version of record not read",true,[93,100,105,112,116,120],{"category":94,"model":95,"canonical":95,"role":96,"dataset":97,"specs":98,"locator":99},"lidar","Velodyne VLP-16","method input","IN2LAAMA UTS datasets (lab, staircase, calibration)","16 channels (plus or minus 15 deg), 10 Hz, 300k points per second, noise plus or minus 3 cm","Sec. VII; Fig. 1",{"category":101,"model":102,"canonical":102,"role":96,"dataset":97,"specs":103,"locator":104},"imu","Xsens MTi-3","100 Hz; noise 0.02 m\u002Fs2 and 0.097 deg\u002Fs; low cost","Sec. VII; Sec. VII-E",{"category":94,"model":106,"canonical":107,"role":108,"dataset":109,"specs":110,"locator":111},"Velodyne HDL-32","Velodyne HDL-32E","dataset sensor","MC2SLAM dataset (campus drive)","about four times the data of the VLP-16; roof of a car","Sec. VII-E-2; Table VII",{"category":101,"model":113,"canonical":113,"role":108,"dataset":109,"specs":114,"locator":115},"Velodyne HDL-32 built-in IMU","not_reported","Sec. VII-E-2",{"category":117,"model":118,"canonical":118,"role":108,"dataset":109,"specs":119,"locator":115},"platform","car","driven around a university campus",{"category":121,"model":122,"canonical":123,"role":124,"dataset":97,"specs":125,"locator":126},"camera","Intel Realsense D435","Intel RealSense D435","compared device","RGB camera used only for the chained IMU-camera-lidar calibration baseline","Sec. VII-F",[],{"totalRows":129,"groupCount":130,"groups":131,"others":422},64,10,[132,264,314,366],{"slug":133,"group":134,"sourceId":5,"sourceLabel":6,"table":135,"selfRows":136,"metrics":137,"seqs":180,"entrants":189,"cells":196,"outcomes":258,"locators":259,"hardware":260,"wordings":261,"notes":262},"in2laama2021-table-i","in2laama2021:Table I","Table I",21,[138,141,145,148,150,152,154,156,158,160,162,164,166,168,170,172,174,176,178],{"label":139,"unit":140,"statistic":114,"alignment":114},"Num. fails; as printed: 0","count",{"label":142,"unit":143,"statistic":144,"alignment":114},"Final pos. error (m); as printed: 0.06 ± 0.03","m","mean",{"label":146,"unit":147,"statistic":144,"alignment":114},"Final rot. error (deg); as printed: 0.12 ± 0.07","deg",{"label":149,"unit":143,"statistic":144,"alignment":114},"Relative pos. error (m), frame-to-frame; as printed: 0.003 ± 3e-4",{"label":151,"unit":147,"statistic":144,"alignment":114},"Relative rot. error (deg), frame-to-frame; as printed: 0.005 ± 1e-4",{"label":153,"unit":143,"statistic":144,"alignment":114},"RMSE pos. error (m); as printed: 0.04 ± 0.02",{"label":155,"unit":147,"statistic":144,"alignment":114},"RMSE rot. error (deg); as printed: 0.09 ± 0.04",{"label":157,"unit":143,"statistic":144,"alignment":114},"Final pos. error (m); as printed: 0.30 ± 0.57",{"label":159,"unit":147,"statistic":144,"alignment":114},"Final rot. error (deg); as printed: 1.37 ± 6.31",{"label":161,"unit":143,"statistic":144,"alignment":114},"Relative pos. error (m), frame-to-frame; as printed: 0.004 ± 0.002",{"label":163,"unit":147,"statistic":144,"alignment":114},"Relative rot. error (deg), frame-to-frame; as printed: 0.009 ± 0.018",{"label":165,"unit":143,"statistic":144,"alignment":114},"RMSE pos. error (m); as printed: 0.19 ± 0.38",{"label":167,"unit":147,"statistic":144,"alignment":114},"RMSE rot. error (deg); as printed: 0.81 ± 3.63",{"label":169,"unit":143,"statistic":144,"alignment":114},"Final pos. error (m); as printed: 0.96 ± 0.61",{"label":171,"unit":147,"statistic":144,"alignment":114},"Final rot. error (deg); as printed: 0.59 ± 3.00",{"label":173,"unit":143,"statistic":144,"alignment":114},"Relative pos. error (m), frame-to-frame; as printed: 0.007 ± 0.003",{"label":175,"unit":147,"statistic":144,"alignment":114},"Relative rot. error (deg), frame-to-frame; as printed: 0.007 ± 0.009",{"label":177,"unit":143,"statistic":144,"alignment":114},"RMSE pos. error (m); as printed: 0.57 ± 0.36",{"label":179,"unit":147,"statistic":144,"alignment":114},"RMSE rot. error (deg); as printed: 0.36 ± 1.71",[181,185,187],{"dataset":182,"sequence":183,"environment":184},"IN2LAAMA simulation (virtual room with 7 planes, VLP-16 and MTi-3 models)","Slow (avg 14.7, max 22.1 deg\u002Fs)","simulated room",{"dataset":182,"sequence":186,"environment":184},"Moderate (avg 49.0, max 78.2 deg\u002Fs)",{"dataset":182,"sequence":188,"environment":184},"Fast (avg 125, max 198 deg\u002Fs)",[190,193,195],{"name":191,"methodId":192,"linkable":91,"proposed":87,"self":87},"[10] (A-LOAM implementation of LOAM)","aloam_software",{"name":194,"methodId":51,"linkable":87,"proposed":87,"self":87},"[5] IN2LAMA (no IMU factors)",{"name":7,"methodId":5,"linkable":91,"proposed":91,"self":91},[197,200,202,204,206,208,211,214,217,220,221,222,225,228,231,233,236,239,240,241,244,247,250,252,255],[198,198,198,198,199,198,199,199,198],0,-1,[201,198,198,198,199,198,199,199,198],1,[203,198,198,198,199,198,199,199,198],2,[203,201,198,205,199,198,199,199,198],0.06,[203,203,198,207,199,198,199,199,198],0.12,[203,209,198,210,199,198,199,199,198],3,0.003,[203,212,198,213,199,198,199,199,198],4,0.005,[203,215,198,216,199,198,199,199,198],5,0.04,[203,218,198,219,199,198,199,199,198],6,0.09,[198,198,201,198,199,198,199,199,198],[203,198,201,198,199,198,199,199,198],[203,223,201,224,199,198,199,199,198],7,0.3,[203,226,201,227,199,198,199,199,198],8,1.37,[203,229,201,230,199,198,199,199,198],9,0.004,[203,130,201,232,199,198,199,199,198],0.009,[203,234,201,235,199,198,199,199,198],11,0.19,[203,237,201,238,199,198,199,199,198],12,0.81,[198,198,203,198,199,198,199,199,198],[203,198,203,198,199,198,199,199,198],[203,242,203,243,199,198,199,199,198],13,0.96,[203,245,203,246,199,198,199,199,198],14,0.59,[203,248,203,249,199,198,199,199,198],15,0.007,[203,251,203,249,199,198,199,199,198],16,[203,253,203,254,199,198,199,199,198],17,0.57,[203,256,203,257,199,198,199,199,198],18,0.36,[],[135],[],[],[263],"Simulated odometry set-up, 50-run Monte Carlo, loop closure off; trajectories average 288.7 m at 4.85 m\u002Fs (max 7.35 m\u002Fs); errors on successful runs only (favours [5] in Fast); values are mean with plus-minus spread",{"slug":265,"group":266,"sourceId":5,"sourceLabel":6,"table":267,"selfRows":229,"metrics":268,"seqs":279,"entrants":288,"cells":290,"outcomes":308,"locators":309,"hardware":310,"wordings":311,"notes":312},"in2laama2021-table-vii","in2laama2021:Table VII","Table VII",[269,272,275,277],{"label":270,"unit":271,"statistic":114,"alignment":85},"Exec. time for the whole sequence (Ne = 100)","s",{"label":273,"unit":274,"statistic":114,"alignment":85},"Mem. data\u002Fprog.","GiB",{"label":276,"unit":274,"statistic":114,"alignment":85},"Mem. optimiser",{"label":278,"unit":271,"statistic":114,"alignment":85},"Exec. time for the whole sequence (Ne = 10)",[280,284,286],{"dataset":281,"sequence":282,"environment":283},"IN2LAAMA UTS datasets and MC2SLAM dataset","Lab (41 s)","offline processing",{"dataset":281,"sequence":285,"environment":283},"Staircase (35 s)",{"dataset":281,"sequence":287,"environment":283},"Outdoor (85 s)",[289],{"name":7,"methodId":5,"linkable":91,"proposed":91,"self":91},[291,293,295,296,298,300,302,304,306],[198,198,198,292,199,198,199,199,198],950,[198,201,198,294,199,198,199,199,198],4.7,[198,203,198,246,199,198,199,199,198],[198,209,201,297,199,198,199,199,198],1306,[198,201,201,299,199,198,199,199,198],4.2,[198,203,201,301,199,198,199,199,198],0.58,[198,198,203,303,199,198,199,199,198],4699,[198,201,203,305,199,198,199,199,198],6.45,[198,203,203,307,199,198,199,199,198],5.77,[],[267],[],[],[313],"Memory consumption and execution time of the offline batch optimisation on real data; Ne frames between optimisations; outdoor uses HDL-32 (about four times the VLP-16 data) and optional ICP loop tests (1588 s of the total)",{"slug":315,"group":316,"sourceId":5,"sourceLabel":6,"table":317,"selfRows":226,"metrics":318,"seqs":335,"entrants":339,"cells":344,"outcomes":360,"locators":361,"hardware":362,"wordings":363,"notes":364},"in2laama2021-table-ii","in2laama2021:Table II","Table II",[319,321,323,325,327,329,331,333],{"label":320,"unit":143,"statistic":144,"alignment":114},"Final pose error (position); as printed: 0.110 m ± 0.043",{"label":322,"unit":147,"statistic":144,"alignment":114},"Final pose error (rotation); as printed: 0.30 ° ± 0.19",{"label":324,"unit":143,"statistic":144,"alignment":114},"RMSE pose error (position); as printed: 0.051 m ± 0.021",{"label":326,"unit":147,"statistic":144,"alignment":114},"RMSE pose error (rotation); as printed: 0.17 ° ± 0.11",{"label":328,"unit":143,"statistic":144,"alignment":114},"Final pose error (position); as printed: 0.011 m ± 0.011",{"label":330,"unit":147,"statistic":144,"alignment":114},"Final pose error (rotation); as printed: 0.15 ° ± 0.12",{"label":332,"unit":143,"statistic":144,"alignment":114},"RMSE pose error (position); as printed: 0.019 m ± 0.007",{"label":334,"unit":147,"statistic":144,"alignment":114},"RMSE pose error (rotation); as printed: 0.10 ° ± 0.07",[336],{"dataset":337,"sequence":338,"environment":184},"IN2LAAMA simulation","closed loops (50 runs)",[340,342],{"name":341,"methodId":5,"linkable":91,"proposed":91,"self":91},"IN2LAAMA (Without loop closure)",{"name":343,"methodId":5,"linkable":91,"proposed":91,"self":91},"IN2LAAMA (With loop closure)",[345,347,348,350,352,354,356,358],[198,198,198,346,199,198,199,199,198],0.11,[198,201,198,224,199,198,199,199,198],[198,203,198,349,199,198,199,199,198],0.051,[198,209,198,351,199,198,199,199,198],0.17,[201,212,198,353,199,198,199,199,198],0.011,[201,215,198,355,199,198,199,199,198],0.15,[201,218,198,357,199,198,199,199,198],0.019,[201,223,198,359,199,198,199,199,198],0.1,[],[317],[],[],[365],"50 simulated closed trajectories (mean 210 m, 3.53 m\u002Fs, 8.16 deg\u002Fs); IN2LAAMA with and without loop closure; mean with plus-minus spread",{"slug":367,"group":368,"sourceId":5,"sourceLabel":6,"table":369,"selfRows":226,"metrics":370,"seqs":387,"entrants":396,"cells":399,"outcomes":416,"locators":417,"hardware":418,"wordings":419,"notes":420},"in2laama2021-table-v","in2laama2021:Table V","Table V",[371,373,375,377,379,381,383,385],{"label":372,"unit":143,"statistic":144,"alignment":85},"calibration translation error; as printed: 10.5e-3 ± 6.34e-3",{"label":374,"unit":147,"statistic":144,"alignment":85},"calibration rotation error; as printed: 0.035 ± 0.037",{"label":376,"unit":143,"statistic":144,"alignment":85},"calibration translation error; as printed: 11.2e-3 ± 6.94e-3",{"label":378,"unit":147,"statistic":144,"alignment":85},"calibration rotation error; as printed: 0.047 ± 0.053",{"label":380,"unit":143,"statistic":144,"alignment":85},"calibration translation error; as printed: 10.8e-3 ± 6.92e-3",{"label":382,"unit":147,"statistic":144,"alignment":85},"calibration rotation error; as printed: 0.062 ± 0.142",{"label":384,"unit":143,"statistic":144,"alignment":85},"calibration translation error; as printed: 16.3e-3 ± 18.7e-3",{"label":386,"unit":147,"statistic":144,"alignment":85},"calibration rotation error; as printed: 0.125 ± 0.326",[388,390,392,394],{"dataset":337,"sequence":389,"environment":184},"initial guess error 0.17 m, 1.74 deg",{"dataset":337,"sequence":391,"environment":184},"initial guess error 0.52 m, 3.47 deg",{"dataset":337,"sequence":393,"environment":184},"initial guess error 0.86 m, 8.55 deg",{"dataset":337,"sequence":395,"environment":184},"initial guess error 1.22 m, 25.1 deg",[397],{"name":398,"methodId":5,"linkable":91,"proposed":91,"self":91},"IN2LAAMA calibration",[400,402,404,406,408,410,412,414],[198,198,198,401,199,198,199,199,198],0.0105,[198,201,198,403,199,198,199,199,198],0.035,[198,203,201,405,199,198,199,199,198],0.0112,[198,209,201,407,199,198,199,199,198],0.047,[198,212,203,409,199,198,199,199,198],0.0108,[198,215,203,411,199,198,199,199,198],0.062,[198,218,209,413,199,198,199,199,198],0.0163,[198,223,209,415,199,198,199,199,198],0.125,[],[369],[],[],[421],"Simulated LiDAR-IMU extrinsic calibration accuracy for four initial-guess error levels; 50 runs of 19.6 s Fast trajectories; mean with plus-minus spread",[423,428,433,439,445,450],{"group":424,"slug":425,"sourceLabel":6,"table":426,"selfRows":218,"datasets":427},"in2laama2021:Table III","in2laama2021-table-iii","Table III",[337],{"group":429,"slug":430,"sourceLabel":6,"table":431,"selfRows":212,"datasets":432},"in2laama2021:Table IV","in2laama2021-table-iv","Table IV",[337],{"group":434,"slug":435,"sourceLabel":6,"table":436,"selfRows":212,"datasets":437},"in2laama2021:Table VI","in2laama2021-table-vi","Table VI",[438],"IN2LAAMA UTS datasets",{"group":440,"slug":441,"sourceLabel":6,"table":442,"selfRows":203,"datasets":443},"in2laama2021:Text Sec.VII-E-2","in2laama2021-text-sec-vii-e-2","Text Sec.VII-E-2",[444],"MC2SLAM dataset",{"group":446,"slug":447,"sourceLabel":6,"table":448,"selfRows":201,"datasets":449},"in2laama2021:Text Sec.VII-C","in2laama2021-text-sec-vii-c","Text Sec.VII-C",[337],{"group":451,"slug":452,"sourceLabel":6,"table":453,"selfRows":201,"datasets":454},"in2laama2021:Text Sec.VII-F","in2laama2021-text-sec-vii-f","Text Sec.VII-F",[438],1790510662843]