FixTheRobot Explore Robotics / Localization & SLAM / Resources

Localization & SLAM · 71 resources

Resources

Where else to learn localization and SLAM: video lecture courses, university syllabi, lecture notes and transcripts, books, the primary papers, software and datasets. Each entry says what it offers, whether it is free, what its licence allows, and how it was checked.

Last checked on 2026-11-12. Back to the track.

Jump to: video lecture courses (15 sources with video), syllabi (13 sources with a public syllabus or schedule), tutorials and textbooks (42 sources with slides, notes or text; 1 with lecture transcripts), papers, software, datasets.

How to read the labels

  • Video Syllabus Notes Transcripts Code Data what the source gives you. Notes means slides, written lecture notes, a tutorial or a book.
  • Free anyone may read or watch it at no cost. It says nothing about copying.
  • Reuse allowed, also commercial the licence lets you copy and adapt the material, with attribution, including in a commercial product.
  • Reuse non-commercial only or same licence required you may copy and adapt it under conditions: not commercially, or only if your result carries the same licence.
  • No reuse the source is under ordinary copyright or says it may not be redistributed. Read it and cite it.
  • Reuse not established no licence statement was found. Assume ordinary copyright: link to it, do not copy it.

What “checked” means. Pages were opened and playlists confirmed automatically, and licences were read where the entry says so. No video was watched and no repository's code was reviewed: the description of each source comes from its own page, title or table of contents. This track cites these sources; its text, figures and code were written for this site and copy none of them.

What the survey found

  • Video lectures are plentiful for filters, grid maps, ICP, pose graphs and multiple view geometry. For LiDAR-inertial systems (LOAM, LIO-SAM, FAST-LIO) and for SLAM on ROS 2 no lecture course was found: the sources are the papers, the repositories and the documentation. Modules 5 and 7 are where this track adds most.
  • Real lecture transcripts are rare. Only MIT 6.801 publishes one for every lecture. Everything else offers slides or written notes; the best written notes are those of MIT VNAV and UCSD ECE276A.
  • Every university syllabus read teaches Lie groups or optimization on manifolds, and most teach factor graphs with incremental smoothing, as lectures of their own. This track mentions both inside other lessons.
  • Only three sources may be adapted on a commercial site with attribution alone: MIT VNAV (CC BY 4.0) and the permissively licensed software and datasets. The rest may be linked and cited.

Start here: one video course and one reading per module

ModuleWatchRead
M1 SLAM FoundationsStachniss: SLAM Course (2013), first lecturesCadena et al., Past, Present, and Future of SLAM
M2 Mathematical FoundationsUMich ROB 530 playlist (estimation, Lie groups)Solà, A micro Lie theory
M3 Mapping and State EstimationStachniss: Mobile Sensing and Robotics 1Thrun, Burgard and Fox, Probabilistic Robotics, chapters 4 to 13 (paid book)
M4 Graph-Based SLAMStachniss: Graph-based SLAM using Pose GraphsDellaert and Kaess, Factor Graphs for Robot Perception
M5 LiDAR SLAMStachniss: Mobile Sensing and Robotics 2 (the three ICP lectures)SLAM Handbook, chapter “LiDAR SLAM”
M6 Visual and Visual-Inertial SLAMCremers: Multiple View GeometryGao et al., Introduction to Visual SLAM
M7 SLAM with ROS 2No lecture course found; use the readingNav2 documentation, with the SLAM Toolbox README and REP 105
M8 Advanced SLAM and ResearchTartan SLAM SeriesSLAM Handbook, parts 2 and 3
M9 CapstoneNo lecture course found; use the readingZhang and Scaramuzza, A Tutorial on Quantitative Trajectory Evaluation

Core references (9)

The sources this track is closest to. If you use only a few, use these: one handbook, one textbook per area, and the one university course whose notes and labs are openly licensed.

  • SLAM Handbook: From Localization and Mapping to Spatial Intelligence

    L. Carlone, A. Kim, T. Barfoot, D. Cremers, F. Dellaert (editors). Cambridge University Press, to appear; public draft released 2024 to 2025 · Book (public draft)

    The current reference, written by the field. Part 1: factor graphs, state representations, robustness to outliers, differentiable optimization, dense maps, certifiable solvers. Part 2: visual, LiDAR, radar, event-based SLAM, inertial and leg odometry. Part 3: deep learning, volume rendering maps, dynamic and deformable scenes, metric-semantic SLAM, spatial AI.

    NotesFreeReuse not established: link onlyModules: M4, M5, M6, M8

    Licence: No licence statement found (the repository has no LICENSE file; the book is to be published by Cambridge University Press). Checked: page opened.

  • Probabilistic Robotics

    S. Thrun, W. Burgard, D. Fox. MIT Press, 2005 · Book

    The standard textbook for the filtering half of the subject: Bayes filters, motion and measurement models, Markov and Monte Carlo localization, occupancy grids, EKF-SLAM, FastSLAM and GraphSLAM.

    NotesPaidNo reuseModules: M1, M2, M3

    Licence: Copyrighted; all rights reserved (standard publisher copyright, not re-read). Checked: could not be reached; listed from prior knowledge.

  • State Estimation for Robotics, 2nd edition

    T. D. Barfoot. Cambridge University Press · Book (author’s PDF)

    The rigorous text for estimation on rotations and poses: linear-Gaussian and nonlinear estimation, handling outliers and biases, three-dimensional geometry, matrix Lie groups, pose estimation and pose-graph problems.

    NotesFreeNo reuseModules: M2, M3, M4

    Licence: Copyrighted; Cambridge University Press is the publisher (title page of the PDF). Checked: document downloaded and read.

  • Factor Graphs for Robot Perception

    F. Dellaert, M. Kaess. Foundations and Trends in Robotics 6(1-2), 2017 · Monograph (author-hosted PDF)

    SLAM as smoothing on a factor graph: sparsity, elimination ordering, incremental smoothing (iSAM), optimization on manifolds and applications. The text behind GTSAM.

    NotesFreeNo reuseModules: M4

    Licence: Copyright 2017 F. Dellaert and M. Kaess (first page of the PDF). Checked: document downloaded and read.

  • Introduction to Visual SLAM: From Theory to Practice

    X. Gao, T. Zhang, Q. Yan, Y. Liu. English edition of “14 Lectures on Visual SLAM”; translation completed 2021 · Book (LaTeX source and PDF) with C++ code

    A complete visual SLAM text with code for every chapter: rigid motion, Lie groups, cameras, nonlinear optimization, visual odometry, back end, loop closure and dense mapping. The same author’s LiDAR book is linked below; its README says the translation is still in progress.

    NotesCodeFreeReuse allowed, same licence requiredModules: M2, M4, M6

    Licence: GPL-3.0 (LICENSE file of the book repository); the code repository slambook2 is MIT. Checked: page opened; licence read.

  • Past, Present, and Future of Simultaneous Localization and Mapping: Toward the Robust-Perception Age

    C. Cadena, L. Carlone, H. Carrillo, Y. Latif, D. Scaramuzza, J. Neira, I. Reid, J. J. Leonard. IEEE Transactions on Robotics, 2016 · Survey paper (arXiv)

    The survey to read first: the standard formulation of SLAM, the anatomy of a system, and the open problems of robustness, scalability, representation and active SLAM as they stood in 2016.

    NotesFreeReuse not established: link onlyModules: M1, M4, M8

    Licence: No licence statement found (the licence field of the arXiv record was not read). Checked: page opened.

  • A Tutorial on Graph-Based SLAM

    G. Grisetti, R. Kümmerle, C. Stachniss, W. Burgard. IEEE Intelligent Transportation Systems Magazine 2(4), 2010 · Tutorial paper

    The derivation that Module 4 follows: the error function of a pose graph, its Jacobians and the Gauss-Newton solution. The authors’ copy is on a university server that did not answer in this session, and the publisher’s page is behind a paywall, so no link is given.

    NotesFreeNo reuseModules: M4

    Licence: Copyright IEEE (not re-read). Checked: could not be reached; listed from prior knowledge.

  • MIT 16.485 Visual Navigation for Autonomous Vehicles (VNAV)

    L. Carlone, MIT · Course: slides, lecture notes, nine labs

    The most complete open university course on the subject without video: 33 lectures from 3D geometry and Lie groups through feature tracking, two-view geometry, RANSAC, nonlinear least squares on manifolds, visual-inertial odometry, place recognition, SLAM formulations, factor graphs and robust estimation, each with slides and many with written notes, plus ROS 2 labs.

    SyllabusNotesCodeFreeReuse allowed, also commercialModules: M2, M4, M6, M7, M8

    Licence: CC BY 4.0 (notice at the foot of every page of vnav.mit.edu); the 2020 edition on MIT OpenCourseWare is CC BY-NC-SA 4.0. Checked: page opened; licence read.

  • Mobile Sensing and Robotics 2 (2021)

    C. Stachniss, University of Bonn · Course: video lectures, slides

    The closest video course to modules 3 to 5: introduction to SLAM, least squares, ICP and point-cloud registration in three parts, graph-based SLAM with pose graphs and with landmarks, hierarchical pose graphs and bundle adjustment, with a short “5 Minutes with Cyrill” video for most topics and all slides as one download.

    VideoSyllabusNotesFreeReuse not established: link onlyModules: M1, M3, M4, M5

    Licence: No licence statement found. Checked: page opened; playlist or video confirmed.

Video lecture courses (14)

Full courses that can be watched in order. For the classical material pick one of the Stachniss or Michigan courses and stay with it; for cameras pick one of the multiple-view-geometry courses.

  • SLAM Course (2013/14, “Robot Mapping”)

    C. Stachniss, University of Freiburg · Video lectures; slides on request

    The classic full course on SLAM: the Kalman filter family, information filters, particle filters, FastSLAM, graph-based approaches, least squares, place recognition and data association.

    VideoSyllabusFreeReuse not established: link onlyModules: M1, M3, M4

    Licence: No licence statement found for the videos; the course page invites other lecturers to use and change the slides with acknowledgement, and sends them by email. Checked: page opened; playlist or video confirmed.

  • Mobile Sensing and Robotics 1 (2021/22)

    C. Stachniss, N. Chebrolu, University of Bonn · Course: video lectures by week

    The localization half: probability primer, Bayes filter, occupancy grid maps, locomotion, motion and observation models, Kalman filter and EKF, EKF localization, particle filter and Monte Carlo localization, then control and planning.

    VideoSyllabusFreeReuse not established: link onlyModules: M1, M2, M3

    Licence: No licence statement found. Checked: page opened; playlist or video confirmed.

  • Online Training: Mobile Robotics

    C. Stachniss, University of Bonn · Ordered list of video lectures

    A self-study path through Stachniss’s lectures, from mathematical background to filters, mapping and graph-based SLAM, in the order he recommends.

    VideoFreeReuse not established: link onlyModules: M1, M2, M3, M4

    Licence: No licence statement found. Checked: page opened; playlist or video confirmed.

  • UMich ROB 530 Mobile Robotics: Methods and Algorithms (Winter 2022)

    M. Ghaffari, University of Michigan · Course: video lectures, slides, notes, homework and code

    A graduate course recorded in full, with seven homework sets: estimation and Kalman filtering, nonlinear filtering, Lie groups and the invariant EKF, localization, mapping and SLAM. Texts: Probabilistic Robotics and Barfoot.

    VideoSyllabusNotesCodeFreeReuse allowed, same licence requiredModules: M2, M3, M4

    Licence: GPL-3.0 (LICENSE file of the repository). Checked: page opened; playlist or video confirmed; licence read.

  • SLAM Lectures

    C. Brenner, Leibniz University Hannover · Video lectures

    A beginner course built around programming exercises in Python (from prior knowledge of the series: motion model, feature extraction from laser scans, Kalman and particle filters, FastSLAM; only the playlist title and channel were confirmed).

    VideoFreeReuse not established: link onlyModules: M1, M3

    Licence: No licence statement found. Checked: playlist or video confirmed.

  • Computer Vision II: Multiple View Geometry

    D. Cremers, Technical University of Munich · Video lectures

    The mathematics of camera-based reconstruction: rigid-body motion and Lie groups, perspective projection, epipolar geometry, the eight-point algorithm, bundle adjustment, direct and dense methods.

    VideoFreeReuse not established: link onlyModules: M2, M6

    Licence: No licence statement found. Checked: page opened; playlist or video confirmed.

  • Photogrammetry I and II (2021/22)

    C. Stachniss, University of Bonn · Course: video lectures, slides as PDF

    Cameras for measurement: image formation, features and matching, camera calibration, the fundamental and essential matrix, triangulation and bundle adjustment. Each lecture has a slide PDF on the page.

    VideoNotesFreeReuse not established: link onlyModules: M6

    Licence: No licence statement found. Checked: page opened; playlist or video confirmed.

  • Computer Vision (University of Tübingen)

    A. Geiger, University of Tübingen · Course: video lectures, slides

    A modern computer vision course whose first half is the geometry visual SLAM needs: image formation, structure from motion, stereo reconstruction.

    VideoNotesFreeReuse not established: link onlyModules: M6

    Licence: No licence statement found. Checked: page opened; playlist or video confirmed.

  • First Principles of Computer Vision

    S. Nayar, Columbia University · Video lecture series

    Short, carefully produced lectures on imaging, features, camera calibration, stereo, optical flow and structure from motion. Good for the camera prerequisites of Module 6.

    VideoFreeReuse not established: link onlyModules: M6

    Licence: No licence statement found. Checked: page opened.

  • MIT 6.801 Machine Vision (Fall 2020)

    B. K. P. Horn. MIT OpenCourseWare · Course: video lectures with transcripts, notes, problem sets

    The one course found that publishes a transcript for every lecture video (PDF and captions). Image formation, motion, photogrammetry, absolute and relative orientation, rotations and quaternions.

    VideoTranscriptsNotesFreeReuse non-commercial onlyModules: M2, M6

    Licence: CC BY-NC-SA 4.0 (licence link on the course page). Checked: page opened; licence read.

  • Self-Driving Cars (University of Bonn)

    C. Stachniss and colleagues, University of Bonn · Video lectures

    Localization and mapping in the setting of an autonomous car (only the playlist title and channel were confirmed).

    VideoFreeReuse not established: link onlyModules: M5, M8

    Licence: No licence statement found. Checked: playlist or video confirmed.

  • Tartan SLAM Series

    AirLab, Carnegie Mellon University · Recorded talks by researchers

    Two seasons of talks by the authors of current systems on what works and what fails in practice: LiDAR and visual-inertial odometry, degraded environments, learning-based methods.

    VideoFreeReuse not established: link onlyModules: M5, M6, M8

    Licence: No licence statement found. Checked: page opened.

  • State Estimation and Localization for Self-Driving Cars

    J. Kelly, S. Waslander, University of Toronto. Coursera · Online course

    Least squares, the Kalman filter family, GNSS and inertial sensing, LiDAR scan matching, and a final project fusing them on a vehicle.

    VideoFree to auditNo reuseModules: M2, M3, M5

    Licence: Coursera terms of use (not re-read). Checked: page opened.

  • Autonomous Mobile Robots

    R. Siegwart and colleagues, ETH Zurich. edX · Online course

    The ETH course behind the textbook Introduction to Autonomous Mobile Robots: locomotion, kinematics, perception, localization, SLAM and planning.

    VideoFree to auditNo reuseModules: M1, M3

    Licence: edX terms of use (not re-read). Checked: page opened.

Course syllabi and lecture notes (8)

University courses that publish what they teach, and in what order, without public video. Useful for comparing this track with how the subject is taught elsewhere.

  • CMU 16-833 Robot Localization and Mapping

    M. Kaess, Carnegie Mellon University · Course page

    The graduate SLAM course of the iSAM author. The public page gives only the description: probabilistic inference for SLAM at the intersection of linear algebra and graphical models. No schedule or slides are posted on it.

    SyllabusFreeReuse not established: link onlyModules: M3, M4

    Licence: No licence statement found. Checked: page opened.

  • UCSD ECE276A Sensing and Estimation in Robotics

    N. Atanasov, UC San Diego · Course: schedule, slides, homework with solutions

    A ten-week schedule with a slide deck and a reading for every topic: optimization, rotations, motion and observation models, factor-graph SLAM, odometry from point features, Bayes and particle filter SLAM, Kalman, EKF and UKF, matrix Lie groups, visual-inertial SLAM, dense mapping.

    SyllabusNotesFreeReuse not established: link onlyModules: M2, M3, M4, M6

    Licence: No licence statement found. Checked: page opened.

  • Vision Algorithms for Mobile Robotics

    D. Scaramuzza, University of Zurich and ETH Zurich · Course: programme, slides, exercises

    Image formation, filtering, feature extraction, multiple view geometry, dense reconstruction, tracking, image retrieval, event-based vision, visual-inertial odometry and SLAM. Slides and exercises are public; the page says the video recordings are on the university platform only.

    SyllabusNotesFreeReuse not established: link onlyModules: M6

    Licence: No licence statement found. Checked: page opened.

  • Imperial College 70003 Advanced Robotics

    S. Leutenegger, Imperial College London · Course description and learning outcomes

    Real-time state estimation and mapping for drones: 6D motion estimation with a camera, fusion with inertial sensors, multi-sensor SLAM with states on a manifold, sparse and dense maps.

    SyllabusFreeReuse not established: link onlyModules: M6, M8

    Licence: No licence statement found. Checked: page opened.

  • Imperial College Robotics course

    A. Davison, Imperial College London · Course: lecture slides, practical sheets

    An undergraduate course in eight lectures with assessed practicals: motion, sensors, probabilistic robotics, Monte Carlo localisation, camera measurements and a final lecture on SLAM.

    SyllabusNotesFreeReuse not established: link onlyModules: M1, M3

    Licence: No licence statement found. Checked: page opened; document downloaded and read.

  • TUM Mobile Robotics (Winter 2023/24)

    S. Leutenegger, Technical University of Munich · Course description

    Perception, estimation, mapping and control for mobile robots, with a substantial part on probabilistic multi-sensor fusion and modern SLAM, including elements of machine learning.

    SyllabusFreeReuse not established: link onlyModules: M3, M6, M8

    Licence: No licence statement found. Checked: page opened.

  • CSCE 774 Robotics Systems (Fall 2023)

    I. Rekleitis, University of South Carolina · Course page with a paper reading list

    A graduate reading course on localization and mapping with an emphasis on visual and inertial data: background lectures, then one paper per session. A model for how to read the primary literature after Module 6.

    SyllabusFreeReuse not established: link onlyModules: M6, M8

    Licence: No licence statement found. Checked: page opened.

  • Stanford AA 275 Navigation for Autonomous Systems

    G. Gao, Stanford University · Course description

    Navigation technology for autonomous systems as taught by a GNSS laboratory. Only the description is public; no lecture videos were found.

    SyllabusFreeReuse not established: link onlyModules: M8

    Licence: No licence statement found. Checked: page opened.

Tutorials and textbooks (11)

Shorter texts for one topic each, and the textbooks for the camera side. Most tutorial papers have an author-hosted copy.

  • A micro Lie theory for state estimation in robotics

    J. Solà, J. Deray, D. Atchuthan. 2018 · Tutorial (arXiv)

    Lie groups with only what estimation needs: the exponential map, plus and minus operators, Jacobians and uncertainty on a manifold, with worked examples for rotations and poses.

    NotesFreeReuse not established: link onlyModules: M2, M4

    Licence: No licence statement found (the licence field of the arXiv record was not read). Checked: page opened.

  • Quaternion kinematics for the error-state Kalman filter

    J. Solà. 2017 · Tutorial (arXiv)

    Quaternions and rotations, conventions and their traps, and the error-state Kalman filter for inertial systems.

    NotesFreeReuse not established: link onlyModules: M2, M6

    Licence: No licence statement found (the licence field of the arXiv record was not read). Checked: page opened.

  • Visual Odometry: Part I and Part II

    D. Scaramuzza, F. Fraundorfer. IEEE Robotics and Automation Magazine, 2011 and 2012 · Tutorial papers (author-hosted PDF)

    The standard two-part tutorial: the first thirty years and fundamentals; then matching, robustness, optimization and applications.

    NotesFreeNo reuseModules: M6

    Licence: Copyright IEEE (not re-read). Checked: document downloaded and read.

  • Visual-Inertial Navigation: A Concise Review

    G. Huang. IEEE International Conference on Robotics and Automation, 2019 · Review paper (arXiv)

    Filtering and optimization approaches to visual-inertial estimation, observability and consistency, calibration and the open problems.

    NotesFreeReuse not established: link onlyModules: M6

    Licence: No licence statement found (the licence field of the arXiv record was not read). Checked: page opened.

  • On-Manifold Preintegration for Real-Time Visual-Inertial Odometry

    C. Forster, L. Carlone, F. Dellaert, D. Scaramuzza. IEEE Transactions on Robotics, 2017 · Paper (arXiv)

    The derivation of IMU preintegration used by LIO-SAM, VINS-Mono and GTSAM.

    NotesFreeReuse not established: link onlyModules: M5, M6

    Licence: No licence statement found (the licence field of the arXiv record was not read). Checked: page opened.

  • iSAM2: Incremental Smoothing and Mapping Using the Bayes Tree

    M. Kaess, H. Johannsson, R. Roberts, V. Ila, J. Leonard, F. Dellaert. International Journal of Robotics Research, 2012 · Paper (author-hosted PDF)

    How a factor graph is updated incrementally instead of being solved again from the start: the back end of many current systems.

    NotesFreeReuse not established: link onlyModules: M4

    Licence: No licence statement found. Checked: document downloaded and read.

  • A Tutorial on Quantitative Trajectory Evaluation for Visual(-Inertial) Odometry

    Z. Zhang, D. Scaramuzza. IEEE/RSJ International Conference on Intelligent Robots and Systems, 2018 · Tutorial paper (author-hosted PDF)

    How to align an estimated trajectory with ground truth and compute absolute and relative errors correctly for each sensor configuration.

    NotesFreeNo reuseModules: M8, M9

    Licence: Copyright IEEE (not re-read). Checked: document downloaded and read.

  • Introduction to Robotics and Perception

    F. Dellaert, S. Hutchinson. Online textbook · Online textbook with notebooks

    An undergraduate text organised by robot, with executable examples that use GTSAM: a gentle route to factor graphs before Module 4.

    NotesCodeFreeReuse not established: link onlyModules: M1, M3, M4

    Licence: No licence statement found (the page says only “Copyright 2023”). Checked: page opened.

  • Computer Vision: Algorithms and Applications, 2nd edition

    R. Szeliski. Springer, 2022 · Book (author’s PDF)

    The broad computer vision reference: image formation, features, alignment, structure from motion and SLAM, depth estimation.

    NotesFreeNo reuseModules: M6

    Licence: The author’s page allows downloading the PDF for personal use. Checked: page opened; licence read.

  • Multiple View Geometry in Computer Vision, 2nd edition

    R. Hartley, A. Zisserman. Cambridge University Press, 2004 · Book

    The reference for projective geometry, two-view and multi-view geometry, and bundle adjustment. The book page has sample chapters and code.

    NotesPaidNo reuseModules: M6

    Licence: Copyrighted; all rights reserved (standard publisher copyright, not re-read). Checked: page opened.

  • Introduction to Autonomous Mobile Robots, 2nd edition

    R. Siegwart, I. Nourbakhsh, D. Scaramuzza. MIT Press, 2011 · Book

    An undergraduate text on locomotion, kinematics, perception, localization and planning. The publisher’s page refused automated access.

    NotesPaidNo reuseModules: M1, M3

    Licence: Copyrighted; all rights reserved (standard publisher copyright, not re-read). Checked: could not be reached; listed from prior knowledge.

Primary papers for the system lessons (6)

The papers behind the algorithm lessons of modules 5, 6 and 8, grouped by family. All are free to read.

  • LOAM: Lidar Odometry and Mapping in Real-time

    J. Zhang, S. Singh. Robotics: Science and Systems, 2014 · Paper, with the papers of its successors

    The paper that defined feature-based LiDAR odometry: edge and plane features, a fast odometry loop and a slow mapping loop. Linked with it: the tightly coupled LiDAR-inertial systems and the registration methods the track discusses.

    NotesFreeReuse not established: link onlyModules: M5

    Licence: No licence statement found. Checked: document downloaded and read; page opened.

  • ORB-SLAM, ORB-SLAM2 and ORB-SLAM3

    R. Mur-Artal, J. M. M. Montiel, J. D. Tardós; C. Campos, R. Elvira, J. J. Gómez Rodríguez. IEEE Transactions on Robotics, 2015, 2017, 2021 · Papers (arXiv)

    The feature-based visual SLAM system in its three generations: monocular; stereo and RGB-D; visual-inertial and multi-map.

    NotesFreeReuse not established: link onlyModules: M6

    Licence: No licence statement found (the licence field of the arXiv record was not read). Checked: page opened.

  • VINS-Mono and Kimera

    T. Qin, P. Li, S. Shen, 2018; A. Rosinol, M. Abate, Y. Chang, L. Carlone, 2020 · Papers (arXiv)

    Two optimization-based visual-inertial systems: a sliding-window monocular estimator with loop closure, and a metric-semantic mapping library built on it.

    NotesFreeReuse not established: link onlyModules: M6, M8

    Licence: No licence statement found (the licence field of the arXiv record was not read). Checked: page opened.

  • Real-Time Loop Closure in 2D LIDAR SLAM (Cartographer) and SLAM Toolbox

    W. Hess, D. Kohler, H. Rapp, D. Andor, 2016; S. Macenski, I. Jambrecic, 2021 · Papers

    The two 2D LiDAR SLAM systems most used with ROS: submaps with branch-and-bound loop closure, and the pose-graph toolbox that Module 7 configures.

    NotesFreeReuse not established: link onlyModules: M3, M7

    Licence: No licence statement found. Checked: page opened.

  • Robust and certifiable estimation: GNC, TEASER and SE-Sync

    H. Yang, P. Antonante, V. Tzoumas, L. Carlone, 2020; H. Yang, J. Shi, L. Carlone, 2020; D. Rosen, L. Carlone, A. Bandeira, J. Leonard, 2019 · Papers (arXiv)

    Graduated non-convexity for outlier rejection, certifiable point-cloud registration, and certifiably correct pose-graph optimization.

    NotesFreeReuse not established: link onlyModules: M4, M8

    Licence: No licence statement found (the licence field of the arXiv record was not read). Checked: page opened.

  • Learned and neural SLAM: DROID-SLAM, iMAP, NICE-SLAM, Gaussian Splatting SLAM, MASt3R-SLAM

    Z. Teed, J. Deng, 2021; E. Sucar et al., 2021; Z. Zhu et al., 2022; H. Matsuki et al., 2024; R. Murai, E. Dexheimer, A. Davison, 2025 · Papers (arXiv)

    The research line of Module 8: learned dense bundle adjustment, implicit neural maps, Gaussian-splat maps, and SLAM built on a learned two-view reconstruction prior.

    NotesFreeReuse not established: link onlyModules: M8

    Licence: No licence statement found (the licence field of the arXiv record was not read). Checked: page opened.

Software and documentation (17)

The libraries and systems the lessons name. Read the documentation before the code; check the licence before you build a product on one.

  • GTSAM

    Georgia Tech Borg Lab · Library and tutorial

    Factor graphs in C++ with Python bindings. The tutorial “Factor Graphs and GTSAM” builds a pose graph, a landmark SLAM problem and a visual-inertial example step by step.

    CodeNotesFreeReuse allowed, also commercialModules: M4, M6

    Licence: Simplified BSD (LICENSE file); bundled third-party libraries carry their own licences. Checked: page opened; licence read.

  • g2o: A General Framework for Graph Optimization

    R. Kümmerle, G. Grisetti, H. Strasdat, K. Konolige, W. Burgard · Library

    The graph optimizer used inside ORB-SLAM and many pose-graph systems.

    CodeFreeReuse allowed, also commercialModules: M4

    Licence: BSD (doc/license-bsd.txt); some optional components are under other licences. Checked: page opened; licence read.

  • Ceres Solver

    Google · Library and tutorial

    A nonlinear least-squares solver. Its tutorial on non-linear least squares covers cost functions, automatic derivatives, robust loss functions and bundle adjustment. The project website is served over plain http and is not linked.

    CodeNotesFreeReuse allowed, also commercialModules: M4, M5, M6

    Licence: BSD 3-Clause (LICENSE file). Checked: licence read.

  • SLAM Toolbox

    S. Macenski · ROS 2 package and README

    The 2D LiDAR SLAM package of Module 7: synchronous and asynchronous mapping, localization, lifelong mapping, serialization of the pose graph, and every parameter explained in the README.

    CodeNotesFreeReuse allowed, same licence requiredModules: M7

    Licence: LGPL-2.1 (LICENSE file). Checked: page opened; licence read.

  • Nav2 documentation

    Open Navigation and the Nav2 contributors · Documentation

    Setup guides for transforms, odometry and sensors, the tutorial on navigating while mapping, and the configuration of AMCL and the map server.

    NotesFreeReuse allowed, also commercialModules: M7

    Licence: Apache-2.0 for files not otherwise marked; some packages are BSD-3-Clause or LGPL-2.1-or-later (LICENSE file of the navigation2 repository). Checked: page opened; licence read.

  • REP 105: Coordinate Frames for Mobile Platforms

    W. Meeussen. ROS Enhancement Proposal · Standard

    The convention every ROS SLAM system assumes: map, odom and base_link, and who publishes which transform. The ros.org page answered with a browser check; the text was read from the REP repository. REP 103 (units and axis conventions) is its companion.

    NotesFreeReuse allowed, also commercialModules: M1, M7

    Licence: Public domain (stated in the document). Checked: licence read.

  • Cartographer

    The Cartographer authors · Library and documentation

    Real-time 2D and 3D SLAM with submaps. The documentation explains the local and global SLAM stages and how to tune them.

    CodeNotesFreeReuse allowed, also commercialModules: M3, M7

    Licence: Apache-2.0 (LICENSE file). Checked: page opened; licence read.

  • KISS-ICP

    I. Vizzo, T. Guadagnino, B. Mersch, L. Wiesmann, J. Behley, C. Stachniss · LiDAR odometry library

    Point-to-point ICP with adaptive thresholding and motion compensation: a small code base to read after the registration lessons.

    CodeFreeReuse allowed, also commercialModules: M5

    Licence: MIT (LICENSE file). Checked: page opened; licence read.

  • LIO-SAM

    T. Shan and colleagues · LiDAR-inertial system

    The reference implementation of the system taught in M5-T8, built on GTSAM.

    CodeFreeReuse allowed, also commercialModules: M5

    Licence: BSD 3-Clause (LICENSE file). Checked: page opened; licence read.

  • FAST-LIO

    HKU MaRS Laboratory · LiDAR-inertial system

    The reference implementation of FAST-LIO and FAST-LIO2: an iterated Kalman filter on raw points with an incremental k-d tree.

    CodeFreeReuse allowed, same licence requiredModules: M5

    Licence: GPL-2.0 (LICENSE file). Checked: page opened; licence read.

  • ORB-SLAM3

    University of Zaragoza SLAM Lab · Visual and visual-inertial SLAM system

    The reference implementation for monocular, stereo, RGB-D and inertial configurations.

    CodeFreeReuse allowed, same licence requiredModules: M6

    Licence: GPL-3.0 (LICENSE file). Checked: page opened; licence read.

  • OpenVINS

    Robot Perception and Navigation Group, University of Delaware · Visual-inertial estimator with documentation

    A filter-based visual-inertial system whose documentation derives the IMU propagation, the update and first-estimate Jacobians: unusually good reading for Module 6.

    CodeNotesFreeReuse allowed, same licence requiredModules: M6

    Licence: GPL-3.0 (stated in the documentation and in the LICENSE file). Checked: page opened; licence read.

  • Kalibr

    Autonomous Systems Lab, ETH Zurich · Calibration toolbox

    Camera intrinsics, multi-camera extrinsics and camera-IMU spatial and temporal calibration: the tool behind M6-T8.

    CodeFreeReuse allowed, also commercialModules: M6

    Licence: BSD-style licence (LICENSE file). Checked: page opened; licence read.

  • RTAB-Map

    M. Labbé, IntRoLab, Université de Sherbrooke · RGB-D, stereo and LiDAR SLAM system

    Appearance-based loop closure with memory management for long-term mapping, with ROS 2 integration.

    CodeNotesFreeReuse allowed, also commercialModules: M6, M7, M8

    Licence: BSD-style licence (LICENSE file). Checked: page opened; licence read.

  • evo

    M. Grupp · Python package

    Trajectory evaluation: absolute and relative pose error, alignment and plots for KITTI, TUM and EuRoC formats and ROS bags.

    CodeFreeReuse allowed, same licence requiredModules: M8, M9

    Licence: GPL-3.0 (LICENSE file). Checked: page opened; licence read.

  • Open3D: ICP registration tutorial

    Open3D contributors · Library tutorial

    Point-to-point and point-to-plane ICP in Python on real scans, with the evaluation of a registration.

    CodeNotesFreeReuse allowed, also commercialModules: M3, M5

    Licence: MIT (LICENSE file). Checked: page opened; licence read.

  • PythonRobotics

    A. Sakai and contributors · Code collection with documentation

    Short Python implementations of EKF and particle filter localization, EKF-SLAM, FastSLAM, ICP matching and graph-based SLAM.

    CodeFreeReuse allowed, also commercialModules: M3, M4

    Licence: MIT (LICENSE file). Checked: page opened; licence read.

Datasets and benchmarks (6)

What published systems are evaluated on. Each has its own trajectory format and its own conditions of use.

  • KITTI odometry benchmark

    A. Geiger, P. Lenz, R. Urtasun. Karlsruhe Institute of Technology and Toyota Technological Institute · Dataset and leaderboard

    Stereo and LiDAR sequences from a car with GPS/INS ground truth, and the relative error metric most LiDAR odometry papers report.

    DataFreeReuse non-commercial onlyModules: M5, M8, M9

    Licence: CC BY-NC-SA 3.0 (stated on the KITTI home page). Checked: page opened; licence read.

  • TUM RGB-D SLAM dataset and benchmark

    J. Sturm and colleagues, Technical University of Munich · Dataset and evaluation tools

    Indoor RGB-D sequences with motion-capture ground truth, and the scripts that defined absolute trajectory error and relative pose error.

    DataCodeFreeReuse allowed, also commercialModules: M6, M8, M9

    Licence: CC BY 4.0 (licence link on the dataset page). Checked: page opened; licence read.

  • EuRoC MAV dataset

    M. Burri and colleagues, ETH Zurich · Dataset

    Stereo and IMU sequences from a small aircraft with accurate ground truth: the standard visual-inertial benchmark. The original page now points to this index of the laboratory’s datasets.

    DataFreeReuse not established: link onlyModules: M6, M8

    Licence: No licence statement found. Checked: page opened.

  • Newer College dataset

    Oxford Robotics Institute · Dataset

    Handheld LiDAR, stereo and inertial data around an Oxford college with a survey-grade ground-truth map.

    DataFreeReuse not established: link onlyModules: M5, M8

    Licence: No licence statement found. Checked: page opened.

  • TartanAir

    AirLab, Carnegie Mellon University · Simulated dataset

    Photo-realistic simulated sequences with hard motion, weather and lighting, made to push visual SLAM to failure.

    DataFreeReuse allowed, also commercialModules: M6, M8

    Licence: CC BY 4.0 (licence link on the dataset page). Checked: page opened; licence read.

  • Hilti SLAM Challenge

    Hilti, with university partners · Dataset and yearly challenge

    Construction-site sequences with millimetre-accurate ground truth and a public leaderboard, renewed each year.

    DataFreeReuse not established: link onlyModules: M8, M9

    Licence: No licence statement found. Checked: page opened.