Map-Merging Algorithms for Visual SLAM: Feasibility Study and Empirical Evaluation

Bokovoy, Andrey; Muraviev, Kirill; Yakovlev, Konstantin

doi:10.1007/978-3-030-59535-7_4

Cited by 9 publications

(7 citation statements)

References 31 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Однако необходимо отметить, что большинство как российских, так и зарубежных исследований связаны именно с технологиями искусственного интеллекта, их разработкой, внедрением и улучшением. Есть работы по отдельным направлениям ИИ, например, можно отметить ряд трудов российских исследователей в сфере ИИ -компьютерное зрение (Khokhlova et al, 2019;Bokovoy et al, 2020), биометрическое распознавание (Vartanov et al, 2020), предиктивная аналитика (Buevich et al, 2021) и др.…”

Section: анализ существующих подходов к классификации технологий в сф...unclassified

Untitled

2022

ESC / ЭиСП

View full text Add to dashboard Cite

Рецензируемый научно-практический журнал, охватывающий вопросы анализа и прогноза изменений в экономике и социальной сфере различных стран и регионов, локальных территорий.Основная цель издания журнала -предоставление широким слоям мировой научной общественности и практическим работникам возможности публиковать результаты изысканий в сфере исследования социально-экономических процессов, знакомиться с различными точками зрения на актуальные проблемы развития экономики и социума, принимать участие в дискуссиях по обсуждаемым темам. В числе основных тем -стратегии развития территорий, региональная и отраслевая экономика, социальное развитие, вопросы формирования доходов бюджетов и рационализации расходов, инновационная экономика, вопросы экономической теории.

show abstract

Section: анализ существующих подходов к классификации технологий в сф...unclassified

Untitled

2022

ESC / ЭиСП

View full text Add to dashboard Cite

show abstract

“…However, some authors have also considered map merging for single-robot applications, where merging is used as a means to facilitate relocalization [8], [3]. This is the use case of map merging on which we focus in this paper.…”

Section: Related Workmentioning

confidence: 99%

“…In the case of graphbased SLAM approaches, the merged gridmap also has the disadvantage that it does not contain any graph structure anymore, so individual lidar scan poses cannot be optimized further when new data becomes available. There is also some research on merging feature maps [5] and point clouds [3]. However, these strategies suffer from similar limitations.…”

Section: Related Workmentioning

confidence: 99%

“…Over the course of its lifetime, it will inevitably happen that localization is lost, e.g, when the robot is kidnapped (turned off, moved by a human, and turned on again), when sensors temporarily provide no or unusable data, or when the SLAM algorithm fails. In such cases, the robot first starts building a new map and later performs relocalization (see [2], [3]) as soon as it can find its pose again with respect to the previously created map. In this case, we merge the previous map with the current map to create a combined map that covers the entire area explored so far.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Detecting Invalid Map Merges in Lifelong SLAM

Holoch¹,

Kurz²,

Biber³

2022

Preprint

View full text Add to dashboard Cite

For Lifelong SLAM, one has to deal with temporary localization failures, e.g., induced by kidnapping. We achieve this by starting a new map and merging it with the previous map as soon as relocalization succeeds. Since relocalization methods are fallible, it can happen that such a merge is invalid, e.g., due to perceptual aliasing. To address this issue, we propose methods to detect and undo invalid merges. These methods compare incoming scans with scans that were previously merged into the current map and consider how well they agree with each other. Evaluation of our methods takes place using a dataset that consists of multiple flat and office environments, as well as the public MIT Stata Center dataset. We show that methods based on a change detection algorithm and on comparison of gridmaps perform well in both environments and can be run in real-time with a reasonable computational cost.

show abstract

“…To run it in monocular mode, we use fully-convolutional neural network (FCNN) like [3] to predict depth of images from camera. Experiments conducted in work [2] show that RTAB-MAP with CNN-predicted depths is able to successfully build a map of indoor scene in most of cases. Average absolute mapping error was about 0.7m, and most of it was the scale error.…”

Section: Slammentioning

confidence: 99%

Enhancing exploration algorithms for navigation with visual SLAM

Muravyev,

Bokovoy,

Yakovlev

2021

Preprint

Self Cite

View full text Add to dashboard Cite

Exploration is an important step in autonomous navigation of robotic systems. In this paper we introduce a series of enhancements for exploration algorithms in order to use them with vision-based simultaneous localization and mapping (vSLAM) methods. We evaluate developed approaches in photo-realistic simulator in two modes: with ground-truth depths and neural network reconstructed depth maps as vSLAM input. We evaluate standard metrics in order to estimate exploration coverage.

show abstract

Map-Merging Algorithms for Visual SLAM: Feasibility Study and Empirical Evaluation

Cited by 9 publications

References 31 publications

Untitled

Untitled

Detecting Invalid Map Merges in Lifelong SLAM

Enhancing exploration algorithms for navigation with visual SLAM

Contact Info

Product

Resources

About