Nonlinear factor recovery for long-term SLAM

  title={Nonlinear factor recovery for long-term SLAM},
  author={Mladen Mazuran and Wolfram Burgard and Gian Diego Tipaldi},
  journal={The International Journal of Robotics Research},
  pages={50 - 72}
For long-term operations, graph-based simultaneous localization and mapping (SLAM) approaches require nodes to be marginalized in order to control the computational cost. In this paper, we present a method to recover a set of nonlinear factors that best represents the marginal distribution in terms of Kullback–Leibler divergence. The proposed method, which we call nonlinear factor recovery (NFR), estimates both the mean and the information matrix of the set of nonlinear factors, where the… 

Approximating Marginalization with Sparse Global Priors for Sliding Window SLAM-Graphs

This work addresses the problem of computing landmark updates live in the vehicle, which requires efficient use of the computational resources, and employs a graph-based sliding window approach for simultaneous localization and incremental map refinement.

Factor descent optimization for sparsification in graph SLAM

Factor Descent is proposed, a new iterative optimization method to sparsify the dense result of node marginalization, which works by iterating factor by factor, which is compared with state-of-the-art methods in real world datasets with regards to the obtained solution and convergence rates.

Graph SLAM Sparsification With Populated Topologies Using Factor Descent Optimization

Factor descent and noncyclic factor descent are proposed, two simple algorithms for SLAM sparsification that match the state-of-art methods without any parameters to be tuned, and are compared against the state of the art with regard to accuracy and CPU time.

Decoupled, consistent node removal and edge sparsification for graph-based SLAM

This paper analytically shows that during marginalization, the correct choice of linearization points in constructing dense marginal factors is to use the relative (local), instead of global, state estimates in the Markov blanket of the marginalized node, which has lacked a general consensus in the literature.

Information Sparsification in Visual-Inertial Odometry

This paper presents a novel approach to tightly couple visual and inertial measurements in a fixed-lag visual-inertial odometry (VIO) framework using information sparsification, which sparsifies the dense prior with a nonlinear factor graph by minimizing the information loss.

A unified resource-constrained framework for graph SLAM

This work attempts to explicitly connect all of the aforementioned resource constraint requirements by considering the node removal and sparsification pipeline in its entirety, and produces node subset selection strategies that are optimal in terms of minimizing the impact of Kullback-Liebler divergence (KLD), of approximating the dense distribution by a sparse one.

A Partial Sparsification Scheme for Visual-Inertial Odometry

  • Z. ZhuWei Wang
  • Computer Science
    2020 IEEE/ASME International Conference on Advanced Intelligent Mechatronics (AIM)
  • 2020
The proposed scheme is to first retain the dense prior from the marginalization excluding visual measurements, followed by a dense marginalization step that connects landmarks, then partially sparsified to extract pseudo factors that maintain the overall sparsity while minimizing the information loss.

Toward Invariant Visual-Inertial State Estimation using Information Sparsification

This work presents a novel approach to tightly couple visual and inertial measurements in a fixed-lag VIO framework using information sparsification and provides detailed derivations of a novel IMU preintegration framework using the group affine properties.

Fast Nonlinear Approximation of Pose Graph Node Marginalization

A fast nonlinear approximation method for marginalizing out nodes on pose graphs for longterm simultaneous localization, mapping, and navigation by re-parameterizing from absolute-to relative-pose spaces, which eschews the expenses of many state-of-the-art convex optimization schemes.



Generic factor-based node marginalization and edge sparsification for pose-graph SLAM

The proposed factor-based method for node marginalization in simultaneous localization and mapping (SLAM) pose-graphs improves upon existing methods in two key ways and outperforms other state-of-the-art methods in terms of Kullback-Leibler divergence.

Nonlinear Graph Sparsification for SLAM

A novel framework for nonlinear graph sparsification in the context of simultaneous localization and mapping is presented, formulated as a convex minimization problem, where the set of nonlinear measurements that best approximate the original distribution are selected.

Generic Node Removal for Factor-Graph SLAM

The proposed GLC method provides a principled and flexible tool to control the computational complexity of long-term graph SLAM, with results shown for 34.9 h of real-world indoor-outdoor data covering 147.4 km collected over 27 mapping sessions spanning a period of 15 months.

iSAM: Fast Incremental Smoothing and Mapping with Efficient Data Association

iSAM is introduced, a novel approach to the problem of simultaneous localization and mapping (SLAM) that addresses the data association problem and allows real-time application in large-scale environments and a QR-factorization of this information matrix is at the heart of this approach.

A Tree Parameterization for Efficiently Computing Maximum Likelihood Maps using Gradient Descent

This paper applies a novel parameterization of the nodes in the graph that significantly improves the performance and enables the algorithm to cope with arbitrary network topologies and converge faster than the other approaches and yields accurate maps of the environment.

Sparse extended information filters: insights into sparsification

It is shown that exaggerated map inconsistency occurs within the global reference frame where estimation is performed, but that empirical testing shows that relative local map relationships are preserved, and a slightly modified version of their sparsification procedure is presented, shown to preserve sparsity while also generating both local and global map estimates comparable to those obtained by the nonsparsified SLAM filter.

Information-theoretic compression of pose graphs for laser-based SLAM

This paper addresses the problem of efficient information-theoretic compression of pose graphs with real-world experiments demonstrating that the approach applies an approximate marginalization technique that is based on Chow–Liu trees and is well suited for long-term mobile robot mapping.

Consistent sparsification for graph optimization

This paper introduces a consistent graph sparsification scheme: sparsifying nodes via marginalization of old nodes, while retaining all the information - which is conveyed in the discarded measurements - about the remaining nodes after marginalization; and sparsify edges by formulating and solving a consistent ℓ1-regularized minimization problem, which automatically promotes the sparsity of the graph.

Square Root SAM: Simultaneous Localization and Mapping via Square Root Information Smoothing

The theory underlying these methods, along with an interpretation of factorization in terms of the graphical model associated with the SLAM problem, are presented, and both simulation results and actual SLAM experiments are presented that underscore the potential of these methods as an alternative to EKF-based approaches.

Conservative edge sparsification for graph SLAM node removal

Results are presented over multiple real-world SLAM graphs and show that the proposed methods enforce a conservative approximation, while achieving low KLD from the true marginalization potential.