[{"data":1,"prerenderedAt":160},["ShallowReactive",2],{"method-hahnel2003_gridfastslam":3},{"method":4,"reference":50,"equipment":72,"figures":91,"results":92},{"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":38,"association":39,"timeModel":40,"deskew":41,"loopClosure":42,"globalOptimization":43,"mapRepresentation":44,"prior":43,"outputGeometry":45,"compute":46,"codeUrl":47,"codeLicense":48,"relatedVersions":49},"hahnel2003_gridfastslam","Hähnel et al., 2003a","Grid-based FastSLAM with scan matching","An efficient FastSLAM algorithm for generating maps of large-scale cyclic environments from raw laser range measurements",2003,"classic","C01","full_slam_with_global_correction","本文把 Rao-Blackwellized 粒子濾波與雷射掃描匹配結合：每 k 步先以前 k-1 筆掃描與最近的里程計讀值做掃描匹配，得到修正後的里程量測並用於粒子取樣，再以第 k 筆掃描計算粒子權重，使每筆資料只使用一次。掃描匹配殘差以三參數誤差模型描述，參數由 Intel Research Lab 資料學得，因此取樣分布比原始里程計集中得多，所需粒子數與重取樣次數下降，也減輕粒子耗盡，使機器人能閉合大迴圈。每個粒子各有一張佔據網格地圖，但只用與其可視區域相交的有限掃描更新。","Grid-based FastSLAM that converts scan-matching results into corrected odometry with a learned error model before Rao-Blackwellized particle sampling, reducing particle count and depletion so that large loops close with about 100 particles.","full_text_reviewed","peer_reviewed_published","background","未在營建場域驗證；實驗在 Intel Research Lab、University of Washington Sieg Hall 等既有建築與模擬環境中進行，只以目視判斷地圖一致性。以掃描匹配先修正里程再進入粒子濾波的設計，是 FastSLAM [fastslam2002] 與 GMapping [gmapping2007] 之間的方法銜接，對大型建築室內 2D 建圖的迴圈閉合穩健性有參考價值（推論）。",[20,21],"completed_building","simulation",[23,24,25,26],"Globally consistent map of the Intel Research Lab (28 m x 28 m, 491 m traveled) built in real time with 100 samples (Sec. IV.A; Fig. 6).","Consistent real-time map of Sieg Hall (50 m x 12 m) despite repeated loops, with 100 samples and a 10 cm grid (Sec. IV.A; Fig. 8).","In simulation, a single-map posterior approach that keeps only the best particle at loop closure produced inconsistencies, while the proposed method produced a consistent map (Sec. IV.B; Fig. 10).","A standard RBPF without scan-matching correction did not converge on the Intel data with up to 1000 samples (Sec. IV.B).",[28,29,30,31],"The real-time 100-sample map is less sharp than the scan-matching-only map; a crisper 500-particle map took several hours (Sec. IV.A).","Each particle's map is updated from a limited set of scans, an approximation (Sec. III).","The scan-matching error model was learned by treating a map produced by the system itself as ground truth (Sec. III).","(inference) Map quality is judged visually; no metric error against an independent reference is given, and the window k is not stated.",[33,34],"2D laser range finder (SICK LMS)","wheel odometry",[36,37],"wheeled UGV (Pioneer 2)","simulation (B21r simulator)","Rao-Blackwellized particle filter over robot paths with one grid map per particle; every k steps a scan-matching-corrected odometry measurement is computed from the k-1 previous scans and the k most recent odometry readings and used for sampling with a learned three-parameter error model, and the k-th scan weights the particles (Sec. III)","grid-based 2D scan matching of a scan against an occupancy grid built from previous measurements, using a beam-endpoint likelihood (for max-range readings the cell 20 cm before the end is assumed free) (Sec. III)","discrete poses","not_reported","implicit through the particle filter; the scan-matching correction reduces resampling operations and particle depletion so that large loops can be closed (Sec. I, III)","none","2D occupancy grid per particle, updated from a limited number of scans that intersect the particle's visible area (constant-time approximation); 10 cm grid in the Sieg Hall run (Sec. III; Sec. IV.A)","2D occupancy grid map","real time with 100 samples; for the standard RBPF, 200 samples was the real-time limit and 1000 samples the memory limit on a 1.8 GHz Pentium IV PC with 768 MB (Sec. IV)",null,"not_applicable (no code linked in the paper)",[],{"id":5,"kind":51,"shortName":7,"title":8,"authors":52,"year":9,"venue":57,"venueType":58,"publisher":59,"volumeIssuePages":60,"doi":61,"arxivId":47,"url":62,"firstPublicDate":63,"publicationStatus":16,"metadataStatus":64,"fulltextStatus":15,"era":10,"classicReason":65,"codeUrl":47,"cluster":11,"topics":66,"mdpi":67,"verification":68,"label":6,"fulltextRoute":69,"versionRead":70,"addedByCensus":71},"method",[53,54,55,56],"Dirk Hähnel","Wolfram Burgard","Dieter Fox","Sebastian Thrun","Proceedings 2003 IEEE\u002FRSJ International Conference on Intelligent Robots and Systems (IROS 2003), Las Vegas, NV","conference","IEEE","vol. 1, pp. 206-211","10.1109\u002Firos.2003.1250629","https:\u002F\u002Fdoi.org\u002F10.1109\u002FIROS.2003.1250629","2003-10","metadata_verified","principle reused: turns scan-matching results into corrected odometry with a learned error model before Rao-Blackwellized particle sampling, cutting particle numbers and depletion; it cites FastSLAM [fastslam2002] and the grid RBPF line later refined by GMapping [gmapping2007] (reviewer synthesis of the lineage).",[11],false,"corrected","NTU institutional (curl)","IEEE Xplore version of record (scanned PDF with OCR text layer, 6 pp.)",true,[73,78,81,87],{"category":74,"model":75,"canonical":75,"role":76,"dataset":47,"specs":41,"locator":77},"platform","Pioneer 2","method input","Sec. IV.A",{"category":79,"model":80,"canonical":80,"role":76,"dataset":47,"specs":41,"locator":77},"lidar","SICK LMS",{"category":82,"model":83,"canonical":83,"role":84,"dataset":47,"specs":85,"locator":86},"compute","1.8GHz Pentium IV PC","compute for runtime","768 MB main memory","Sec. IV.B",{"category":88,"model":89,"canonical":89,"role":76,"dataset":47,"specs":90,"locator":86},"other","B21r simulator","simulator of a B21r robot used to generate the Wean Hall data (32 m x 10 m, 251 m, noise added to the ground truth)",[],{"totalRows":93,"groupCount":94,"groups":95,"others":159},4,1,[96],{"slug":97,"group":98,"sourceId":5,"sourceLabel":6,"table":99,"selfRows":93,"metrics":100,"seqs":110,"entrants":121,"cells":129,"outcomes":141,"locators":146,"hardware":151,"wordings":152,"notes":153},"hahnel2003-gridfastslam-text-sec-iv","hahnel2003_gridfastslam:Text Sec. IV","Text Sec. IV",[101,104,107],{"label":102,"unit":103,"statistic":41,"alignment":43},"number of samples used for the real-time map","samples",{"label":105,"unit":106,"statistic":41,"alignment":43},"number of particles for the offline map","particles",{"label":108,"unit":109,"statistic":41,"alignment":43},"loop closure outcome","not_applicable",[111,115,118],{"dataset":112,"sequence":113,"environment":114},"Intel Research Lab (Pioneer 2, SICK LMS)","Intel Research Lab log","indoor building",{"dataset":116,"sequence":117,"environment":114},"University of Washington Sieg Hall","Sieg Hall log",{"dataset":119,"sequence":120,"environment":21},"B21r simulator, Wean Hall","simulated run",[122,124,126],{"name":123,"methodId":5,"linkable":71,"proposed":71,"self":71},"proposed RBPF with scan-matching-corrected odometry",{"name":125,"methodId":5,"linkable":71,"proposed":71,"self":71},"proposed RBPF (offline)",{"name":127,"methodId":128,"linkable":71,"proposed":67,"self":67},"particle filter strategy of Thrun et al. [20], [19] (single map)","thrun2000_3dmapping",[130,134,136,138,140],[131,131,131,132,131,131,133,133,131],0,100,-1,[94,94,131,135,94,94,133,133,94],500,[131,131,94,132,131,137,133,133,137],2,[137,137,137,47,137,139,133,133,139],3,[131,137,137,47,139,139,133,133,93],[142,143,144,145],"consistent map in real time","computation took several hours","failed: inconsistencies after closing the loop","consistent map",[147,148,149,150],"Sec. IV.A; Fig. 6","Sec. IV.A; Fig. 9","Sec. IV.A; Fig. 8","Sec. IV.B; Fig. 10",[],[],[154,155,156,157,158],"Intel Research Lab (28 m x 28 m, 491 m traveled); map judged globally consistent and built in real time","Intel Research Lab, offline map with more particles; crisper than the 100-sample map","Sieg Hall fourth floor (50 m x 12 m), several loops, 10 cm grid","Simulated Wean Hall (32 m x 10 m, 251 m, noise added); single-map posterior approach keeping only the best particle at loop closure","Simulated Wean Hall, proposed method",[],1790510661426]