Factor Graphs and GTSAM: A Hands-on Introduction

By Frank Dellaert. 2012

842 citations recorded by the source corpus.

View Paper: https://www.semanticscholar.org/paper/b94fbf48299d78cd586c057e700763ec09b88f80

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

Citations in the corpus

Visual-Inertial Navigation Algorithm Development Using Photorealistic Camera Simulation in the Loop (2018, methodology): “(3) is solved using the incremental smoothing algorithm iSAM2 [35] implemented in the GTSAM 4.0 toolbox [36].”
Past, Present, and Future of Simultaneous Localization and Mapping: Toward the Robust-Perception Age (2016, unknown): “Current SLAM libraries (e.g., GTSAM [83], g2o [164], Ceres [3], iSAM [146], and SLAM++ [229]) are able to solve problems with tens of thousands of variables in few seconds.”
Kimera: an Open-Source Library for Real-Time Metric-Semantic Localization and Mapping (2019, methodology): “2) Robust PGO: This module is implemented in GTSAM, and includes a modern outlier rejection method, Incremental Consistent Measurement Set Maximization (PCM) [46], that we tailor to a single-robot and online setup.”
Legged Robot State-Estimation Through Combined Forward Kinematic and Preintegrated Contact Factors (2017, methodology): “We implemented both factors in GTSAM [22] using iSAM2 as the solver [23].”
On-Manifold Preintegration for Real-Time Visual--Inertial Odometry (2015, background): “0 optimization toolbox [7] and provide additional theoretical derivations and implementation details in the Appendix of this paper.”
HDMapNet: An Online HD Map Construction and Evaluation Framework (2021, background): “Second, estimating accurate poses of ego vehicle is formulated as a non-linear least-square problem [6] or a factor graph [7] which is critical to build a globally consistent map.”
OpenVINS: A Research Platform for Visual-Inertial Estimation (2020, methodology): “Inspired by graph-based optimization frameworks such as GTSAM [32], we abstract away from the user the need to directly manipulate the covariance and instead provide the tools to automatically manage the state and its covariance.”
Global Pose Estimation with an Attention-Based Recurrent Network (2018, background): “Several such solvers achieve excellent performance, for example, g2o [25], TSAM [8], Ceres, iSAM [20], SLAM++ [38], and recently [2] for optimization with semantic data association.”
AutoRally: An Open Platform for Aggressive Autonomous Driving (2019, methodology): “The iSAM2 algorithm is a part of the Georgia Tech smoothing and mapping (GTSAM) [34], [35] software package, which uses a factor graph representation to iteratively solve the smoothing and mapping problem.”
Kimera: From SLAM to spatial perception with 3D dynamic scene graphs (2021, methodology): “Finally, states that fall out of the smoothing horizon are marginalized out using GTSAM.”
Probabilistic data association for semantic SLAM (2017, methodology): “We implemented our algorithm in C++ using GTSAM [39] and its iSAM2 implementation as the optimization back-end.”
A Benchmark Comparison of Monocular Visual-Inertial Odometry Algorithms for Flying Robots (2018, background): “0 optimization toolbox7 [35], are publicly available, however the integration of these into a single system is not currently public.”
SE-Sync: A certifiably correct algorithm for synchronization over the special Euclidean group (2016, methodology): “Our experimental implementation of SE-Sync 16 is written in C++ (Rosen and Carlone, 2017), and we compare it against the Powell’s Dog-Leg implementation supplied by GTSAM,17 a highly optimized, state-of-the-art software library specifically designed for large-scale SLAM and bundle adjustment applica”
RTAB‐Map as an open‐source lidar and visual simultaneous localization and mapping library for large‐scale and long‐term online operation (2018, methodology): “From the online documentation (which is the only documentation available), loop closures seem detected locally by scan matching when the robot visits previous locations, to then optimize the map using Georgia tech smoothing and mapping (GTSAM) (Dellaert, 2012).”
Factor Graphs for Robot Perception (2017, methodology): “A simple but computationally expensive way to calculate Ga R0 is numerical di erentiation, but more advanced code-bases use symbolic derivatives or some type of automatic di erentiation [113, 47].”