[{"data":1,"prerenderedAt":338},["ShallowReactive",2],{"method-zhu2022liinit":3},{"method":4,"reference":53,"equipment":73,"figures":105,"results":106},{"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":21,"limitations":27,"sensors":33,"platform":36,"estimator":38,"association":39,"timeModel":40,"deskew":41,"loopClosure":42,"globalOptimization":42,"mapRepresentation":43,"prior":44,"outputGeometry":45,"compute":46,"codeUrl":47,"codeLicense":48,"relatedVersions":49},"zhu2022liinit","Zhu et al., 2022b","LI-Init","Robust Real-time LiDAR-inertial Initialization",2022,"recent","C13","sensing_calibration_sync_preprocessing","LI-Init 在 LiDAR 慣性里程計啟動前，自動判斷資料激勵是否足夠，並線上估計 LiDAR 與 IMU 的時間偏移、外參、重力向量與 IMU 偏差。時間偏移先以互相關粗估，再與旋轉外參聯合最佳化。作者強調若時間偏移未知，依賴 IMU 的運動畸變補償就無法正確執行。","Online initialization estimating LiDAR-IMU time offset, extrinsics, gravity and biases with excitation checking, feeding FAST-LIO2.","full_text_reviewed","peer_reviewed_published","main_body","未在工地測試；自組手持設備在建築量測常見，未同步時的時間偏移估計直接影響去畸變（推論）。",[20],"controlled_experiment",[22,23,24,25,26],"With artificially shifted IMU timestamps (0.05 to 0.5 s), time-offset RMSE about 1.6 to 3.4 ms on Livox sensors (Table II)","Extrinsic translation error about 0.6 to 1.3 cm mean and rotation error below 1 deg against CAD reference across three LiDARs (Table III)","On two PandarXT sequences (first 40 s): rotation error 0.6208 deg versus 1.0375 deg (LI-Calib) and 0.8483 deg (Target-Free); translation 0.0162 m versus diverged (LI-Calib) and 0.0187 m (Target-Free); 10.2 s versus 332.6 s and 115.7 s (Table IV)","FAST-LIO2 with the LI-Init time offset: end-to-end drift 0.0102 m over an 11.36 m handheld loop, versus 0.246 m with FAST-LIO2 internal synchronization (Sec. IV-A, Fig. 6)","No calibration target, extra sensor, hardware synchronization or stationary start required; an excitation assessment tells the user how to move the device (Sec. I, Sec. II, Sec. III-C.5)",[28,29,30,31,32],"Assumes an unknown but constant time offset (Sec. I contributions, Sec. III-C.2)","Extrinsic error larger for the spinning PandarXT than the Livox units, because frame splitting shrinks the sub-frame FoV so the device must move slowly, lowering IMU excitation and SNR (Table III, Sec. IV-B.1)","Motion-based: needs sufficient excitation, judged by a singular-value threshold, and LiDAR odometry degeneration prevents data from being used (Sec. III-A, Sec. III-C.5)","Ground-truth time offset is unavailable for unsynchronized rigs; temporal accuracy is shown with artificial offsets on factory-synchronized Livox units and indirectly through LIO drift (Sec. IV-A)","Extrinsic accuracy is validated indirectly through a CAD-designed relative pose between two Pixhawk mounting poses; the baseline comparison uses only two PandarXT sequences (Sec. IV-B)",[34,35],"3D LiDAR: Livox Avia (small FoV), Livox Mid360 (non-repetitive scanning), Hesai PandarXT (mechanical spinning); all set to 10 Hz","6-axis IMU Bosch BMI088, both inside the Livox LiDARs (factory hardware-synchronized) and inside a Pixhawk flight controller (unsynchronized with PandarXT); raw data 200 Hz",[37],"handheld","LiDAR odometry with error-state iterated Kalman filter, then cross-correlation temporal alignment and joint temporal-spatial optimization","LiDAR-only odometry uses scan-to-map point-to-plane residuals (modified from FAST-LIO2), chosen over NDT scan-to-scan so that non-repetitive LiDARs are supported; LiDAR and IMU data are associated by cross-correlation of angular-velocity magnitudes, then least-squares alignment of angular velocity and acceleration","Constant angular and linear velocity model between scans, with each input frame split into sub-frames to reduce model mismatch; IMU data linearly interpolated to LiDAR-odometry timestamps; the unknown constant time offset is estimated as an integer number of odometry intervals by cross-correlation and refined by a sub-interval residual using a constant angular acceleration approximation","During initialization, points are deskewed without the IMU: each point is projected to the scan-end frame using the constant-velocity prediction, because the unsynchronized IMU cannot yet be used; Fig. 3 compares Mid360 maps with and without this compensation; after initialization FAST-LIO2 uses the calibrated offset","none","Point map used for scan-to-map registration inside the FAST-LIO2-derived LiDAR odometry; the data structure is not described in this paper","none (initial extrinsic set to identity in tests)","time offset, extrinsic, gravity, IMU bias as initial states for FAST-LIO2","Desktop Intel i7-10700 at 2.90 GHz with 32 GB RAM; LiDAR odometry about 8 ms per sub-frame; initialization solver below 500 ms once data are collected; 10.2 s total calibration on 40 s of PandarXT data versus 332.6 s for LI-Calib and 115.7 s for Target-Free","https:\u002F\u002Fgithub.com\u002Fhku-mars\u002FLiDAR_IMU_Init","GPL-2.0",[50],{"relation":51,"title":8,"doi_or_url":52},"preprint","https:\u002F\u002Farxiv.org\u002Fabs\u002F2202.11006",{"id":5,"kind":54,"shortName":7,"title":8,"authors":55,"year":9,"venue":59,"venueType":60,"publisher":61,"volumeIssuePages":62,"doi":63,"arxivId":64,"url":52,"firstPublicDate":65,"publicationStatus":16,"metadataStatus":66,"fulltextStatus":15,"era":10,"classicReason":67,"codeUrl":47,"cluster":11,"topics":68,"mdpi":69,"verification":70,"label":6,"fulltextRoute":71,"versionRead":72,"addedByCensus":69},"method",[56,57,58],"Fangcheng Zhu","Yunfan Ren","Fu Zhang","2022 IEEE\u002FRSJ International Conference on Intelligent Robots and Systems (IROS)","conference","IEEE","pp. 3948-3955","10.1109\u002Firos47612.2022.9982225","2202.11006","2022-02-22","metadata_verified","not_applicable",[11],false,"confirmed","arXiv","arXiv 2202.11006v5 (2022-09-15), header 'Accepted for IROS 2022'; not the IEEE version of record",[74,81,86,90,95,100],{"category":75,"model":76,"canonical":76,"role":77,"dataset":78,"specs":79,"locator":80},"lidar","Livox Avia","method input",null,"small FoV, non-repetitive scanning; output set to 10 Hz; built-in BMI088 IMU hardware-synchronized in factory","Fig. 1, Sec. IV, Sec. IV-A, Table II to III",{"category":75,"model":82,"canonical":83,"role":77,"dataset":78,"specs":84,"locator":85},"Livox Mid360","Livox MID-360","non-repetitive scanning; output set to 10 Hz; built-in BMI088 IMU hardware-synchronized in factory","Fig. 1, Sec. IV, Sec. IV-A, Table II to III, Fig. 3, Fig. 8",{"category":75,"model":87,"canonical":87,"role":77,"dataset":78,"specs":88,"locator":89},"Hesai PandarXT","mechanical spinning; output set to 10 Hz; not synchronized with the Pixhawk IMU","Fig. 1, Sec. IV, Sec. IV-A to IV-C, Table III to IV, Fig. 9",{"category":91,"model":92,"canonical":92,"role":77,"dataset":78,"specs":93,"locator":94},"imu","Bosch BMI088","6-axis IMU inside both the Pixhawk flight controller and the Livox LiDARs; raw data 200 Hz","Sec. IV (setup)",{"category":96,"model":97,"canonical":97,"role":77,"dataset":78,"specs":98,"locator":99},"other","Pixhawk flight controller","houses the BMI088 IMU; fixed at two poses I1 and I2 with CAD relative pose of 0 deg rotation and 0.25 m translation (CAD accuracy stated as 0.01 deg and millimetres)","Fig. 1, Sec. IV, Sec. IV-B.1",{"category":101,"model":102,"canonical":102,"role":103,"dataset":78,"specs":104,"locator":94},"compute","Intel i7-10700","compute for runtime","desktop computer, 2.90 GHz, 32 GB RAM; all experiments",[],{"totalRows":107,"groupCount":108,"groups":109,"others":331},36,5,[110,193,251,300],{"slug":111,"group":112,"sourceId":5,"sourceLabel":6,"table":113,"selfRows":114,"metrics":115,"seqs":127,"entrants":140,"cells":144,"outcomes":187,"locators":188,"hardware":189,"wordings":190,"notes":191},"zhu2022liinit-table-ii","zhu2022liinit:Table II","Table II",18,[116,120,123],{"label":117,"unit":118,"statistic":119,"alignment":42},"calibrated time offset Mean[s]","s","mean",{"label":121,"unit":118,"statistic":122,"alignment":42},"time offset RMSE[s]","RMSE",{"label":124,"unit":125,"statistic":126,"alignment":42},"NEES (defined in caption as RMSE divided by the set offset)","%","not_reported",[128,132,134,136,138,139],{"dataset":129,"sequence":130,"environment":131},"authors' handheld sequences (Livox Avia, built-in IMU)","artificial offset 0.05 s","laboratory scene",{"dataset":129,"sequence":133,"environment":131},"artificial offset 0.1 s",{"dataset":129,"sequence":135,"environment":131},"artificial offset 0.5 s",{"dataset":137,"sequence":130,"environment":131},"authors' handheld sequences (Livox Mid360, built-in IMU)",{"dataset":137,"sequence":133,"environment":131},{"dataset":137,"sequence":135,"environment":131},[141],{"name":142,"methodId":5,"linkable":143,"proposed":143,"self":143},"LI-Init (proposed)",true,[145,149,152,155,157,159,161,163,165,167,170,172,174,177,179,181,183,185],[146,146,146,147,148,146,148,148,146],0,0.049,-1,[146,150,146,151,148,146,148,148,146],1,0.0016,[146,153,146,154,148,146,148,148,146],2,3.2,[146,146,150,156,148,146,148,148,146],0.0988,[146,150,150,158,148,146,148,148,146],0.0017,[146,153,150,160,148,146,148,148,146],1.7,[146,146,153,162,148,146,148,148,146],0.4989,[146,150,153,164,148,146,148,148,146],0.0018,[146,153,153,166,148,146,148,148,146],0.36,[146,146,168,169,148,146,148,148,146],3,0.0479,[146,150,168,171,148,146,148,148,146],0.0034,[146,153,168,173,148,146,148,148,146],6.8,[146,146,175,176,148,146,148,148,146],4,0.0982,[146,150,175,178,148,146,148,148,146],0.0028,[146,153,175,180,148,146,148,148,146],2.8,[146,146,108,182,148,146,148,148,146],0.4984,[146,150,108,184,148,146,148,148,146],0.0029,[146,153,108,186,148,146,148,148,146],0.58,[],[113],[],[],[192],"Temporal initialization with artificial IMU timestamp shifts on Livox LiDARs with built-in IMUs; 5 laboratory sequences per LiDAR; true offset about 0 s (factory sync) plus the added shift",{"slug":194,"group":195,"sourceId":5,"sourceLabel":6,"table":196,"selfRows":197,"metrics":198,"seqs":210,"entrants":218,"cells":220,"outcomes":245,"locators":246,"hardware":247,"wordings":248,"notes":249},"zhu2022liinit-table-iii","zhu2022liinit:Table III","Table III",12,[199,202,205,208],{"label":200,"unit":201,"statistic":119,"alignment":42},"relative error Rot(deg), mean","deg",{"label":203,"unit":201,"statistic":204,"alignment":42},"relative error Rot(deg), SD","std",{"label":206,"unit":207,"statistic":119,"alignment":42},"relative error Trans(m), mean","m",{"label":209,"unit":207,"statistic":204,"alignment":42},"relative error Trans(m), SD",[211,214,216],{"dataset":212,"sequence":213,"environment":126},"authors' handheld sequences (Pixhawk IMU at poses I1 and I2)","5 sequences, Livox Mid360",{"dataset":212,"sequence":215,"environment":126},"5 sequences, Livox Avia",{"dataset":212,"sequence":217,"environment":126},"5 sequences, Hesai PandarXT",[219],{"name":142,"methodId":5,"linkable":143,"proposed":143,"self":143},[221,223,225,227,229,231,233,235,237,239,241,243],[146,146,146,222,148,146,148,148,146],0.2472,[146,150,146,224,148,146,148,148,146],0.2043,[146,153,146,226,148,146,148,148,146],0.0081,[146,168,146,228,148,146,148,148,146],0.0075,[146,146,150,230,148,146,148,148,146],0.4019,[146,150,150,232,148,146,148,148,146],0.1708,[146,153,150,234,148,146,148,148,146],0.0064,[146,168,150,236,148,146,148,148,146],0.0069,[146,146,153,238,148,146,148,148,146],0.7244,[146,150,153,240,148,146,148,148,146],0.5076,[146,153,153,242,148,146,148,148,146],0.0133,[146,168,153,244,148,146,148,148,146],0.0102,[],[196],[],[],[250],"Extrinsic initialization: absolute error of the relative pose between two Pixhawk IMU mounting poses (from two LI-Init calibrations) against the CAD design (0 deg, 0.25 m); mean and SD over 5 sequences per LiDAR; initial extrinsic set to identity",{"slug":252,"group":253,"sourceId":5,"sourceLabel":6,"table":254,"selfRows":168,"metrics":255,"seqs":262,"entrants":267,"cells":275,"outcomes":291,"locators":293,"hardware":295,"wordings":297,"notes":298},"zhu2022liinit-table-iv","zhu2022liinit:Table IV","Table IV",[256,258,260],{"label":257,"unit":201,"statistic":119,"alignment":42},"Rotation(deg)",{"label":259,"unit":207,"statistic":119,"alignment":42},"Translation(m)",{"label":261,"unit":118,"statistic":126,"alignment":67},"Time(s), total calibration time",[263],{"dataset":264,"sequence":265,"environment":266},"two PandarXT plus Pixhawk sequences from Sec. IV-B.1","first 40 s of each sequence","calibration scene of Fig. 7(e) (not described further)",[268,270,273],{"name":269,"methodId":5,"linkable":143,"proposed":143,"self":143},"Proposed",{"name":271,"methodId":272,"linkable":143,"proposed":69,"self":69},"LI-Calib [14]","lv2020licalib",{"name":274,"methodId":78,"linkable":69,"proposed":69,"self":69},"Target-Free [15]",[276,278,280,282,283,285,287,289],[146,146,146,277,148,146,148,148,146],0.6208,[146,150,146,279,148,146,148,148,146],0.0162,[146,153,146,281,148,146,146,148,146],10.2,[150,150,146,78,146,150,148,148,146],[150,153,146,284,148,146,146,148,146],332.6,[153,146,146,286,148,146,148,148,146],0.8483,[153,150,146,288,148,146,148,148,146],0.0187,[153,153,146,290,148,146,146,148,146],115.7,[292],"diverged (refinement fails; authors attribute it to missing gravity initialization)",[254,294],"Table IV and note",[296],"desktop Intel i7-10700 @2.90 GHz, 32 GB RAM",[],[299],"Extrinsic calibration comparison on two Hesai PandarXT plus Pixhawk sequences, first 40 s (400 scans) each; average relative IMU pose error against CAD; time offset pre-compensated for the baselines; default baseline parameters; all on the i7-10700 desktop",{"slug":301,"group":302,"sourceId":5,"sourceLabel":6,"table":303,"selfRows":153,"metrics":304,"seqs":310,"entrants":312,"cells":317,"outcomes":322,"locators":325,"hardware":327,"wordings":328,"notes":329},"zhu2022liinit-text-sec-iv-c","zhu2022liinit:Text Sec.IV-C","Text Sec.IV-C",[305,308],{"label":306,"unit":307,"statistic":119,"alignment":67},"average processing time of a sub-frame (LiDAR odometry)","ms",{"label":309,"unit":307,"statistic":126,"alignment":67},"total time of initialization solver (pre-processing, temporal, extrinsic and gravity initialization) once data are collected",[311],{"dataset":126,"sequence":126,"environment":126},[313,315],{"name":314,"methodId":5,"linkable":143,"proposed":143,"self":143},"LI-Init LiDAR odometry",{"name":316,"methodId":5,"linkable":143,"proposed":143,"self":143},"LI-Init initialization solver",[318,320],[146,146,146,319,146,146,146,148,146],8,[150,150,146,321,150,146,146,148,146],500,[323,324],"other: stated as 'about 8 ms'","other: reported as an upper bound (less than 500 ms)",[326],"Sec. IV-C",[296],[],[330],"Runtime statements in the time consumption evaluation; desktop Intel i7-10700",[332],{"group":333,"slug":334,"sourceLabel":6,"table":335,"selfRows":150,"datasets":336},"zhu2022liinit:Text Sec.IV-A","zhu2022liinit-text-sec-iv-a","Text Sec.IV-A",[337],"authors' handheld sequence (Hesai PandarXT plus Pixhawk IMU)",1790510658388]