[{"data":1,"prerenderedAt":166},["ShallowReactive",2],{"method-fastslam2_2003":3},{"method":4,"reference":48,"equipment":68,"figures":94,"results":95},{"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":32,"platform":34,"estimator":36,"association":37,"timeModel":38,"deskew":39,"loopClosure":40,"globalOptimization":41,"mapRepresentation":42,"prior":41,"outputGeometry":43,"compute":44,"codeUrl":45,"codeLicense":46,"relatedVersions":47},"fastslam2_2003","Montemerlo et al., 2003","FastSLAM 2.0","FastSLAM 2.0: An Improved Particle Filtering Algorithm for Simultaneous Localization and Mapping that Provably Converges",2003,"classic","C01","full_slam_with_global_correction","FastSLAM 2.0 修改原 FastSLAM 的取樣方式，在抽樣機器人位姿時同時考慮最新量測，而不只依賴運動模型。作者證明對線性高斯 SLAM，在所有特徵被無限次觀測且已知一個特徵位置的條件下，單一粒子即可在期望值意義上收斂到正確地圖（Sec. 5）。在 Victoria Park 公開資料上，作者報告其精度明顯優於原版 FastSLAM。","Adds measurement-informed pose proposals to FastSLAM and proves single-particle convergence for linear-Gaussian SLAM, with accuracy gains reported on Victoria Park.","full_text_reviewed","peer_reviewed_published","supplementary","not_reported（僅使用 Victoria Park 公園的戶外車輛資料與模擬資料，與營建場域無關）",[20,21,22],"public_benchmark","simulation","independent_reference",[24,25,26,27,28],"Convergence proof for linear SLAM with M = 1 particle, described by the authors as the first for a constant-time SLAM algorithm (Sec. 5","Sec. 7) | Abstract reports an order-of-magnitude accuracy improvement over the original FastSLAM on real-world data","Fig. 2 shows FastSLAM 2.0 performing about equally well for any particle number while regular FastSLAM performs poorly with very small particle sets (abstract","Sec. 6) | With M = 1 on Victoria Park, accuracy previously reached only by O(N^2) EKF-style methods or FastSLAM with M = 50","processing took 54 s, under 4% of the 1,550 s acquisition time (Sec. 6)",[30,31],"Authors state multiple particles remain warranted when data association is ambiguous (Sec. 7).","The proof applies only to a restricted linear Gaussian SLAM class (Sec. 5).",[33],"[\"range finder (type not stated in the paper)\",\"vehicle odometry (described as relatively inaccurate)\",\"GNSS (DGPS used for evaluation only)\"]",[35,21],"vehicle","Rao-Blackwellized particle filter with pose proposal conditioned on the current measurement","per-particle maximum-likelihood association that accounts for the sampled pose (Sec. 4.4); new features created when the measurement probability falls below a threshold, and spurious features removed by a log-odds existence filter (Sec. 4.5)","discrete poses","not_reported","implicit; no explicit module","none","point landmarks (per-particle EKFs)","landmark map and vehicle path","each update takes constant time with M = 1 (Sec. 5); on a 1 GHz Pentium PC the Victoria Park data (1,550 s acquisition) took 54 s with FastSLAM 2.0 (M = 1), 315 s with regular FastSLAM (M = 50) and 7,807 s with an EKF (Sec. 6 table)",null,"not_applicable",[],{"id":5,"kind":49,"shortName":7,"title":8,"authors":50,"year":9,"venue":55,"venueType":56,"publisher":57,"volumeIssuePages":58,"doi":45,"arxivId":45,"url":59,"firstPublicDate":60,"publicationStatus":16,"metadataStatus":61,"fulltextStatus":15,"era":10,"classicReason":62,"codeUrl":45,"cluster":11,"topics":63,"mdpi":64,"verification":65,"label":6,"fulltextRoute":66,"versionRead":67,"addedByCensus":64},"method",[51,52,53,54],"Michael Montemerlo","Sebastian Thrun","Daphne Koller","Ben Wegbreit","Proceedings of the Eighteenth International Joint Conference on Artificial Intelligence (IJCAI-03)","conference","IJCAI (proceedings published by Morgan Kaufmann; hosted at ijcai.org)","pp. 1151-1156","https:\u002F\u002Fwww.ijcai.org\u002FProceedings\u002F03\u002FPapers\u002F165.pdf","2003","metadata_verified","necessary technical node: shows that incorporating the latest observation into the pose proposal changes particle-filter SLAM behaviour, an idea that GMapping's improved proposal distribution also pursues.",[11],false,"corrected","publisher OA","IJCAI-03 proceedings paper, pp. 1151-1156, PDF from ijcai.org (scanned pages with an ABBYY OCR text layer); no DOI exists",[69,75,82,86,90],{"category":70,"model":71,"canonical":71,"role":72,"dataset":45,"specs":73,"locator":74},"compute","1GHz Pentium PC","compute for runtime","1 GHz","Sec. 6 (runtime table)",{"category":76,"model":77,"canonical":77,"role":78,"dataset":79,"specs":80,"locator":81},"gnss","differential GPS","reference or ground truth","Victoria Park (Sydney)","used for evaluation only","Sec. 6",{"category":83,"model":84,"canonical":84,"role":85,"dataset":79,"specs":39,"locator":81},"platform","outdoor vehicle","dataset sensor",{"category":87,"model":88,"canonical":88,"role":85,"dataset":79,"specs":89,"locator":81},"other","range finder","low-noise; used for landmark detection (type not stated)",{"category":87,"model":91,"canonical":91,"role":85,"dataset":79,"specs":92,"locator":93},"odometry","described as relatively inaccurate; raw-odometry average RMS error 93.6 m","Sec. 6, Fig. 1a",[],{"totalRows":96,"groupCount":97,"groups":98,"others":165},3,2,[99,133],{"slug":100,"group":101,"sourceId":5,"sourceLabel":6,"table":102,"selfRows":97,"metrics":103,"seqs":107,"entrants":111,"cells":117,"outcomes":125,"locators":126,"hardware":129,"wordings":130,"notes":131},"fastslam2-2003-text-sec-6","fastslam2_2003:Text Sec. 6","Text Sec. 6",[104],{"label":105,"unit":106,"statistic":39,"alignment":41},"number of landmarks in the map","landmarks",[108],{"dataset":79,"sequence":109,"environment":110},"full data set (3.5 km)","outdoor park",[112,115],{"name":113,"methodId":5,"linkable":114,"proposed":114,"self":114},"FastSLAM 2.0, M = 1, without feature management",true,{"name":116,"methodId":5,"linkable":114,"proposed":114,"self":114},"FastSLAM 2.0, M = 1, with feature management",[118,122],[119,119,119,120,121,119,121,121,119],0,768,-1,[123,119,119,124,121,123,121,121,119],1,343,[],[127,128],"Sec. 6, Fig. 1b","Sec. 6, Fig. 1c",[],[],[132],"Number of landmarks in the Victoria Park map with M = 1, without and with the feature-management rule of Sec. 4.5",{"slug":134,"group":135,"sourceId":5,"sourceLabel":6,"table":136,"selfRows":123,"metrics":137,"seqs":141,"entrants":143,"cells":151,"outcomes":158,"locators":159,"hardware":161,"wordings":162,"notes":163},"fastslam2-2003-table-sec-6","fastslam2_2003:Table Sec. 6","Table Sec. 6",[138],{"label":139,"unit":140,"statistic":39,"alignment":41},"processing time for the whole data set","s",[142],{"dataset":79,"sequence":109,"environment":110},[144,146,149],{"name":145,"methodId":45,"linkable":64,"proposed":64,"self":64},"EKF",{"name":147,"methodId":148,"linkable":114,"proposed":64,"self":64},"regular FastSLAM, M=50 particles","fastslam2002",{"name":150,"methodId":5,"linkable":114,"proposed":114,"self":114},"FastSLAM 2.0, M=1 particle",[152,154,156],[119,119,119,153,121,119,119,121,119],7807,[123,119,119,155,121,119,119,121,119],315,[97,119,119,157,121,119,119,121,119],54,[],[160],"Sec. 6 table",[71],[],[164],"Total time to process the Victoria Park data set on a 1 GHz Pentium PC; data acquisition took 1,550 s",[],1790510662680]