[{"data":1,"prerenderedAt":351},["ShallowReactive",2],{"method-rflio2021":3},{"method":4,"reference":56,"equipment":79,"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":22,"limitations":28,"sensors":34,"platform":37,"estimator":39,"association":40,"timeModel":41,"deskew":42,"loopClosure":43,"globalOptimization":44,"mapRepresentation":45,"prior":46,"outputGeometry":47,"compute":48,"codeUrl":49,"codeLicense":50,"relatedVersions":51},"rflio2021","Qian et al., 2021","RF-LIO","RF-LIO: Removal-First Tightly-coupled Lidar Inertial Odometry in High Dynamic Environments",2021,"recent","C05","full_slam_with_global_correction","RF-LIO 以 LIO-SAM 為基礎，處理大量移動物體時「先要準確位姿才能移除動態點、但動態點又破壞配準」的循環問題。新關鍵影格到達時先不做掃描配準，而是以 IMU 預積分取得初始位姿，並依預測的平移與旋轉誤差決定距離影像的角解析度；將目前掃描與周邊特徵子地圖投影成同解析度距離影像，依可見性差異移除子地圖中的動態點，再做 LOAM 特徵配準。若以邊緣點距離計算的收斂分數未達門檻，就以新解析度重複移除與配準；收斂並完成圖最佳化後，再以細解析度移除目前關鍵影格殘留的動態點。","LIO-SAM-based LIO for highly dynamic scenes that removes moving points before scan matching: the IMU-predicted pose error sets the resolution of scan and submap range images, visibility differences flag dynamic points, and removal and LOAM-feature matching are repeated at new resolutions until an edge-distance convergence score is met.","full_text_reviewed","peer_reviewed_published","supplementary","論文未在營建場域評估；自建 Urban 資料的路線雖經過施工區，但未單獨分析（Sec. IV-D）。其先移除動態點再配準的做法，概念上可對應工地上移動的工人、車輛與機具造成的重影與配準偏差，對需要靜態點雲地圖的施工記錄有參考價值；但在開闊或被大型機具遮擋視野的工地，作者指出的兩項限制可能直接出現（推論）。",[20,21],"public_benchmark","cross_site",[23,24,25,26,27],"Average moving-point removal rate of 96.1% relative to LIO-SAM maps on three self-collected datasets (Table III)","On UrbanLoco high-dynamic sequences, ATE RMSE 15.89 and 12.17 m versus 62.43 and 36.98 m for LIO-SAM and 203.78 and 175.69 m for LOAM (Table V)","Removal-first variants also improved ATE on low and medium dynamic data, e.g. Urban 6.46 m versus 10.21 m for LIO-SAM (Table IV)","Removal before matching reduced per-scan runtime compared with removal after matching (Table VI)","No training data or semantic labels required (Sec. V)",[29,30,31,32,33],"In very open scenes without far points behind moving objects, the visibility test cannot remove moving points (Sec. V)","Not suitable when moving objects fully block the sensor field of view (Sec. V)","Points close to the ground (below 0.5 m) and returns from beams parallel to the ground are not removed (Sec. IV-C)","Runtime exceeds the 100 ms LiDAR period on Suburban and CARussianHill (112 and 121 ms) (Table VI)","Ground truth is GPS only; the removal rate is measured relative to LIO-SAM's residual moving points rather than labelled ground truth (Sec. IV-A, IV-C)",[35,36],"3D LiDAR at 10 Hz (model not named)","IMU at 400 Hz (model not named)",[38],"not_reported (self-collected Urban, Campus and Suburban datasets and UrbanLoco CA sequences; carrier not described in the paper)","LIO-SAM factor graph in GTSAM (IMU preintegration, LiDAR odometry and loop closure factors) with an added removal-first loop: IMU prior, dynamic-point removal, scan matching, convergence check and repeated removal at a new resolution (Sec. III-A, IV-A)","LOAM edge and planar features with nearest-neighbour point-to-line and point-to-plane distances to a feature submap, after removing submap points flagged as dynamic (Sec. III-E)","discrete keyframes with IMU preintegration (Sec. III-B)","IMU-based motion compensation of each scan as in LIO-SAM (Sec. III-A, Fig. 2)","yes; same loop detection as LIO-SAM (Sec. IV-A)","factor-graph optimization with GTSAM as in LIO-SAM (Sec. IV-A)","keyframe-based edge and planar feature map from which moving points are removed; global point map without ghost tracks (Fig. 5)","none","trajectory and a static point cloud map with moving-object points removed","per-scan runtime of RF-LIO (FA) 61 to 121 ms on an Intel i7-10700K CPU; below 100 ms on low and medium dynamic data except Suburban (112 ms) (Table VI, Sec. IV-A)",null,"not_applicable (no public code found)",[52],{"relation":53,"title":54,"doi_or_url":55},"preprint","arXiv 2206.09463 v1 (2022-06-19), posted after the conference","https:\u002F\u002Farxiv.org\u002Fabs\u002F2206.09463",{"id":5,"kind":57,"shortName":7,"title":8,"authors":58,"year":9,"venue":63,"venueType":64,"publisher":65,"volumeIssuePages":66,"doi":67,"arxivId":68,"url":69,"firstPublicDate":70,"publicationStatus":16,"metadataStatus":71,"fulltextStatus":15,"era":10,"classicReason":72,"codeUrl":49,"cluster":11,"topics":73,"mdpi":74,"verification":75,"label":6,"fulltextRoute":76,"versionRead":77,"addedByCensus":78},"method",[59,60,61,62],"Chenglong Qian","Zhaohong Xiang","Zhuoran Wu","Hongbin Sun","2021 IEEE\u002FRSJ International Conference on Intelligent Robots and Systems (IROS)","conference","IEEE","pp. 4421-4428","10.1109\u002Firos51168.2021.9636624","2206.09463","https:\u002F\u002Fdoi.org\u002F10.1109\u002FIROS51168.2021.9636624","2021-09-27","metadata_verified","not_applicable",[11],false,"corrected","arXiv","arXiv v1 (2022-06-19; only version, posted after IROS 2021); IEEE version of record not read",true,[80,87,94,100],{"category":81,"model":82,"canonical":82,"role":83,"dataset":84,"specs":85,"locator":86},"gnss","GPS (receiver model not stated)","reference or ground truth","self-collected Urban, Campus, Suburban and UrbanLoco","used only as ground truth","Sec. IV-A",{"category":88,"model":89,"canonical":90,"role":91,"dataset":49,"specs":92,"locator":93},"compute","Intel i7-10700k (as written)","Intel i7-10700k","compute for runtime","ROS on Ubuntu Linux","Sec. IV-A; Table VI",{"category":95,"model":96,"canonical":96,"role":97,"dataset":49,"specs":98,"locator":99},"lidar","3D LiDAR (model not stated)","method input","LiDAR scans at 10 Hz (Fig. 2)","Fig. 2; Sec. IV-A",{"category":101,"model":102,"canonical":102,"role":97,"dataset":49,"specs":103,"locator":104},"imu","IMU (model not stated)","IMU data at 400 Hz (Fig. 2)","Fig. 2; Sec. III-B; Sec. IV-A",[],{"totalRows":107,"groupCount":108,"groups":109,"others":350},38,4,[110,182,247,303],{"slug":111,"group":112,"sourceId":5,"sourceLabel":6,"table":113,"selfRows":114,"metrics":115,"seqs":120,"entrants":133,"cells":140,"outcomes":175,"locators":176,"hardware":177,"wordings":179,"notes":180},"rflio2021-table-vi","rflio2021:Table VI","Table VI",15,[116],{"label":117,"unit":118,"statistic":119,"alignment":72},"runtime for processing one scan","ms","not_reported",[121,125,127,129,131],{"dataset":122,"sequence":123,"environment":124},"self-collected datasets and UrbanLoco","Urban","per scan",{"dataset":122,"sequence":126,"environment":124},"Campus",{"dataset":122,"sequence":128,"environment":124},"Suburban",{"dataset":122,"sequence":130,"environment":124},"CAMarketStreet",{"dataset":122,"sequence":132,"environment":124},"CARussianHill",[134,136,138],{"name":135,"methodId":5,"linkable":78,"proposed":74,"self":78},"RF-LIO (After)",{"name":137,"methodId":5,"linkable":78,"proposed":78,"self":78},"RF-LIO (First)",{"name":139,"methodId":5,"linkable":78,"proposed":78,"self":78},"RF-LIO (FA)",[141,145,148,151,154,156,158,160,161,163,165,167,169,171,173],[142,142,142,143,144,142,142,144,142],0,86,-1,[142,142,146,147,144,142,142,144,142],1,97,[142,142,149,150,144,142,142,144,142],2,118,[142,142,152,153,144,142,142,144,142],3,105,[142,142,108,155,144,142,142,144,142],134,[146,142,142,157,144,142,142,144,142],55,[146,142,146,159,144,142,142,144,142],68,[146,142,149,147,144,142,142,144,142],[146,142,152,162,144,142,142,144,142],93,[146,142,108,164,144,142,142,144,142],107,[149,142,142,166,144,142,142,144,142],61,[149,142,146,168,144,142,142,144,142],74,[149,142,149,170,144,142,142,144,142],112,[149,142,152,172,144,142,142,144,142],96,[149,142,108,174,144,142,142,144,142],121,[],[113],[178],"Intel i7-10700K",[],[181],"Runtime of RF-LIO variants for processing one scan",{"slug":183,"group":184,"sourceId":5,"sourceLabel":6,"table":185,"selfRows":186,"metrics":187,"seqs":192,"entrants":201,"cells":211,"outcomes":240,"locators":242,"hardware":243,"wordings":244,"notes":245},"rflio2021-table-iv","rflio2021:Table IV","Table IV",9,[188],{"label":189,"unit":190,"statistic":191,"alignment":119},"absolute trajectory RMSE","m","RMSE",[193,197,199],{"dataset":194,"sequence":195,"environment":196},"self-collected datasets","Urban (6390.33 m, low dynamic)","outdoor urban, campus and suburban routes",{"dataset":194,"sequence":198,"environment":196},"Campus (1007.97 m, medium dynamic)",{"dataset":194,"sequence":200,"environment":196},"Suburban (1890.44 m, medium dynamic)",[202,205,208,209,210],{"name":203,"methodId":204,"linkable":78,"proposed":74,"self":74},"LOAM","loam2014",{"name":206,"methodId":207,"linkable":78,"proposed":74,"self":74},"LIO-SAM","liosam2020",{"name":135,"methodId":5,"linkable":78,"proposed":74,"self":78},{"name":137,"methodId":5,"linkable":78,"proposed":78,"self":78},{"name":139,"methodId":5,"linkable":78,"proposed":78,"self":78},[212,214,216,217,219,221,223,225,226,228,230,232,234,236,238],[142,142,142,213,144,142,144,144,142],244.19,[142,142,146,215,144,142,144,144,142],118.49,[142,142,149,49,142,142,144,144,142],[146,142,142,218,144,142,144,144,142],10.21,[146,142,146,220,144,142,144,144,142],0.66,[146,142,149,222,144,142,144,144,142],1.53,[149,142,142,224,144,142,144,144,142],10.79,[149,142,146,220,144,142,144,144,142],[149,142,149,227,144,142,144,144,142],1.51,[152,142,142,229,144,142,144,144,142],7.72,[152,142,146,231,144,142,144,144,142],0.64,[152,142,149,233,144,142,144,144,142],1.42,[108,142,142,235,144,142,144,144,142],6.46,[108,142,146,237,144,142,144,144,142],0.61,[108,142,149,239,144,142,144,144,142],1.25,[241],"failed",[185],[],[],[246],"Low and medium dynamic self-collected datasets; LiDAR and IMU only, GPS as ground truth; RF-LIO and LIO-SAM share feature extraction and loop closure",{"slug":248,"group":249,"sourceId":5,"sourceLabel":6,"table":250,"selfRows":251,"metrics":252,"seqs":259,"entrants":269,"cells":272,"outcomes":297,"locators":298,"hardware":299,"wordings":300,"notes":301},"rflio2021-table-iii","rflio2021:Table III","Table III",8,[253,256],{"label":254,"unit":255,"statistic":119,"alignment":72},"residual moving object points in map","points",{"label":257,"unit":258,"statistic":119,"alignment":72},"removal rate of moving object points","%",[260,262,264,266],{"dataset":194,"sequence":123,"environment":261},"urban terrain incl. residential area, overpass and construction area",{"dataset":194,"sequence":126,"environment":263},"XJTU campus with pedestrians",{"dataset":194,"sequence":128,"environment":265},"suburban road with moving vehicles",{"dataset":194,"sequence":267,"environment":268},"Average","average of three datasets",[270,271],{"name":206,"methodId":207,"linkable":78,"proposed":74,"self":74},{"name":7,"methodId":5,"linkable":78,"proposed":78,"self":78},[273,275,277,279,281,283,285,287,289,291,293,295],[142,142,142,274,144,142,144,144,142],85890,[146,142,142,276,144,142,144,144,142],1803,[146,146,142,278,144,142,144,144,142],97.9,[142,142,146,280,144,142,144,144,142],134092,[146,142,146,282,144,142,144,144,142],5766,[146,146,146,284,144,142,144,144,142],95.7,[142,142,149,286,144,142,144,144,142],198113,[146,142,149,288,144,142,144,144,142],8898,[146,146,149,290,144,142,144,144,142],95.5,[142,142,152,292,144,142,144,144,142],139365,[146,142,152,294,144,142,144,144,142],5489,[146,146,152,296,144,142,144,144,142],96.1,[],[250],[],[],[302],"Residual moving-object points counted in the maps of LIO-SAM and RF-LIO (same feature extraction); removal rate relative to LIO-SAM",{"slug":304,"group":305,"sourceId":5,"sourceLabel":6,"table":306,"selfRows":307,"metrics":308,"seqs":310,"entrants":317,"cells":323,"outcomes":344,"locators":345,"hardware":346,"wordings":347,"notes":348},"rflio2021-table-v","rflio2021:Table V","Table V",6,[309],{"label":189,"unit":190,"statistic":191,"alignment":119},[311,315],{"dataset":312,"sequence":313,"environment":314},"UrbanLoco","CAMarketStreet (5690.98 m)","highly urbanized streets with many moving objects",{"dataset":312,"sequence":316,"environment":314},"CARussianHill (3570.38 m)",[318,319,320,321,322],{"name":203,"methodId":204,"linkable":78,"proposed":74,"self":74},{"name":206,"methodId":207,"linkable":78,"proposed":74,"self":74},{"name":135,"methodId":5,"linkable":78,"proposed":74,"self":78},{"name":137,"methodId":5,"linkable":78,"proposed":78,"self":78},{"name":139,"methodId":5,"linkable":78,"proposed":78,"self":78},[324,326,328,330,332,334,336,338,340,342],[142,142,142,325,144,142,144,144,142],203.78,[142,142,146,327,144,142,144,144,142],175.69,[146,142,142,329,144,142,144,144,142],62.43,[146,142,146,331,144,142,144,144,142],36.98,[149,142,142,333,144,142,144,144,142],23.98,[149,142,146,335,144,142,144,144,142],12.91,[152,142,142,337,144,142,144,144,142],15.83,[152,142,146,339,144,142,144,144,142],12.44,[108,142,142,341,144,142,144,144,142],15.89,[108,142,146,343,144,142,144,144,142],12.17,[],[306],[],[],[349],"High dynamic UrbanLoco sequences with many moving objects; LiDAR and IMU only, GPS as ground truth",[],1790510665743]