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

Kretzschmar, Henrik; Stachniss, Cyrill

doi:10.1177/0278364912455072

Cited by 103 publications

(115 citation statements)

References 27 publications

Supporting

Mentioning

110

Contrasting

Unclassified

Order By: Relevance

“…(1) Expected information gain The expected information gain [16,17] is defined as the prediction of uncertainty of the robot which can be reduced by the observation of a candidate node. The expected amount of information in which an observation contributes to the belief, can be measured by the entropy.…”

Section: Generation Of Nodes Based On Entropy-driven Strategymentioning

confidence: 99%

Map segmentation for simultaneous localization and mapping in ruins

Wang

et al. 2016

Advanced Robotics

View full text Add to dashboard Cite

The distribution of environmental features of the internal ruins which are formed by a randomly seismic disaster is unpredictable. Therefore, the existing methods of map segmentation, which need to preset parameters, cannot be directly used. Considering the lack of prior knowledge, a map segmentation method based on the spectral clustering is proposed in the framework of hierarchical simultaneous localization and mapping (SLAM) algorithm. The method solves the problem of incremental complexity of SLAM algorithm using the division of environment. In accordance with the similarity of observed environment, a weighted graph is established. The nodes in the graph are generated by measuring the expected information gain and position redundancy. Then, the graph is partitioned into subjective results of map segment based on the criterion of minimum normalized cut. On the basis of the inherent sparse of SLAM, the proposed algorithm not only reduces the cost of calculation, but also minimizes the loss of information in order to ensure the global consistency. Finally, the feasibility and effectiveness of the algorithm are verified by simulation and experiment. ARTICLE HISTORY

show abstract

Section: Generation Of Nodes Based On Entropy-driven Strategymentioning

confidence: 99%

Map segmentation for simultaneous localization and mapping in ruins

Wang

et al. 2016

Advanced Robotics

View full text Add to dashboard Cite

show abstract

“…Keyframe-based approaches reduce the number of constraints in visual SLAM by keeping only a subset of measurement frames [14]. Kretzschmar et al [15] estimated the mutual information of laser-scan measurements with regard to an occupancy grid, only incorporating new measurements when they are sufficiently informative. They marginalize out old poses to bound map growth to a fixed memory size.…”

Section: Related Workmentioning

confidence: 99%

Location utility-based map reduction

Steiner¹,

Huang

Leonard³

2015

2015 IEEE International Conference on Robotics and Automation (ICRA)

View full text Add to dashboard Cite

Abstract-Maps used for navigation often include a database of location descriptions for place recognition (loop closing), which permits bounded-error performance. A standard posegraph SLAM system adds a new entry for every new pose into the location database, which grows linearly and unbounded in time and thus becomes unsustainable. To address this issue, in this paper we propose a new map-reduction approach that pre-constructs a fixed-size place-recognition database amenable to the limited storage and processing resources of the vehicle by exploiting the high-level structure of the environment as well as the vehicle motion. In particular, we introduce the concept of location utility -which encapsulates the visitation probability of a location and its spatial distribution relative to nearby locations in the database -as a measure of the value of potential loop-closure events to occur at that location. While finding the optimal reduced location database is NPhard, we develop an efficient greedy algorithm to sort all the locations in a map based on their relative utility without access to sensor measurements or the vehicle trajectory. This enables pre-determination of a generic, limited-size place-recognition database containing the N best locations in the environment. To validate the proposed approach, we develop an open-source street-map simulator using real city-map data and show that an accurate map (pose-graph) can be attained even when using a place-recognition database with only 1% of the entries of the corresponding full database.

show abstract

“…In several applications, it is difficult to solve the optimization problem (9) or to compute the integral (8). Therefore, one rather looks for a lower bound for (8), and iteratively optimizes this lower bound.…”

Section: Upper Bounds Approximationsmentioning

confidence: 99%

“…Therefore, one rather looks for a lower bound for (8), and iteratively optimizes this lower bound. This idea is an old one [26].…”

Section: Upper Bounds Approximationsmentioning

confidence: 99%

“…In [7], Ila et al consider the problem of pose graph optimization and propose information-theoretic measures to select only the most informative edges, so to prevent uncontrolled growth of the graph. In [8], Stachniss and Kretzschmar propose a graph compression technique for the specific case of laserbased SLAM. In [9], Huang et al focus on consistency issues and propose a node marginalization scheme, based on Schur complement; moreover, they introduce a technique for edge sparsification, based on 1 -regularized optimization.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Eliminating conditionally independent sets in factor graphs: A unifying perspective based on smart factors

Carlone

Kira

Beall

et al. 2014

2014 IEEE International Conference on Robotics and Automation (ICRA)

View full text Add to dashboard Cite

Abstract-Factor graphs are a general estimation framework that has been widely used in computer vision and robotics. In several classes of problems a natural partition arises among variables involved in the estimation. A subset of the variables are actually of interest for the user: we call those target variables. The remaining variables are essential for the formulation of the optimization problem underlying maximum a posteriori (MAP) estimation; however these variables, that we call support variables, are not strictly required as output of the estimation problem. In this paper, we propose a systematic way to abstract support variables, defining optimization problems that are only defined over the set of target variables. This abstraction naturally leads to the definition of smart factors, which correspond to constraints among target variables. We show that this perspective unifies the treatment of heterogeneous problems, ranging from structureless bundle adjustment to robust estimation in SLAM. Moreover, it enables to exploit the underlying structure of the optimization problem and the treatment of degenerate instances, enhancing both computational efficiency and robustness.

show abstract

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

Cited by 103 publications

References 27 publications

Map segmentation for simultaneous localization and mapping in ruins

Map segmentation for simultaneous localization and mapping in ruins

Location utility-based map reduction

Eliminating conditionally independent sets in factor graphs: A unifying perspective based on smart factors

Contact Info

Product

Resources

About