Article
Engineering, Ocean
Ling Yu, Li Ye, Ma Teng, Cong Zheng, Xu Shuo, Zhihui Li
Summary: This paper proposes an active bathymetric SLAM method that actively detects loops by driving an AUV into areas with significant topographic relief. It introduces a terrain-complexity-related mutual information utility function to optimize the trade-off between exploration and revisiting tasks. Simulation experiments demonstrate that the proposed method outperforms state-of-the-art algorithms.
APPLIED OCEAN RESEARCH
(2023)
Article
Robotics
Tiziano Guadagnino, Luca Di Giammarino, Giorgio Grisetti
Summary: This letter presents a novel hierarchical algorithm HiPE for pose graph initialization, which utilizes a sparse graph construction and maximum likelihood estimates to achieve non-linear initialization and guide the fine-grained optimization process of the final solution. Experimental results show that HiPE leads to a more efficient and robust optimization process compared to existing methods.
IEEE ROBOTICS AND AUTOMATION LETTERS
(2022)
Article
Computer Science, Software Engineering
Zheng-Jun Du, Shi-Sheng Huang, Tai-Jiang Mu, Qunhe Zhao, Ralph R. Martin, Kun Xu
Summary: Accurate camera pose tracking in dynamic environments is achieved in this article through a novel RGB-D SLAM approach, which utilizes long-term observations and conditional random fields for more accurate dynamic 3D landmark detection. Evaluation results demonstrate the superiority of this method over existing approaches.
IEEE TRANSACTIONS ON VISUALIZATION AND COMPUTER GRAPHICS
(2022)
Article
Computer Science, Artificial Intelligence
Luying Zhong, Zhaoliang Chen, Zhihao Wu, Shide Du, Zheyi Chen, Shiping Wang
Summary: This article proposes a learnable GCN-based framework to obtain optimal graph structures and designs dual-GCN-based meta-channels to explore local and global relations. The introduction of SGIB maximizes the mutual information between the same and different meta-channels, improving node classification performance.
IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS
(2023)
Article
Robotics
Jiahui Fu, Chengyuan Lin, Yuichi Taguchi, Andrea Cohen, Yifu Zhang, Stephen Mylabathula, John J. Leonard
Summary: The ability to process environment maps across multiple sessions is crucial for robots operating over extended periods. This letter explores the problem of change detection using a novel map representation called Plane Signed Distance Fields (PlaneSDF). The proposed approach, which involves comparing height maps and conducting three-dimensional geometric validation using SDF features, is shown to be effective for detecting changed objects.
IEEE ROBOTICS AND AUTOMATION LETTERS
(2022)
Article
Mathematics
Xuyan Xiang, Jieming Zhou
Summary: Long-term memory behavior is an important phenomenon in time series analysis. The excess entropy approach is used to classify long-term and short-term memory in stationary time series. Simulation results demonstrate the effectiveness of the approach on various stochastic sequences. The approach has advantages over traditional methods as it is invariant under instantaneous one-to-one transformation and has weak moment conditions.
Article
Computer Science, Theory & Methods
FatimaEzzahra Laghrissi, Samira Douzi, Khadija Douzi, Badr Hssina
Summary: An Intrusion Detection System (IDS) is a device or software application that monitors networks for malicious activities, with deep learning algorithms proving effective in improving detection efficiency. In this study, PCA and Mutual information were utilized as dimensionality reduction and feature selection techniques for a deep learning attack detection model based on Long Short-Term Memory (LSTM). The experimental results on the KDD99 benchmark dataset demonstrated that PCA-based models achieved the highest accuracy in both training and testing for binary and multiclass classification.
JOURNAL OF BIG DATA
(2021)
Article
Robotics
Yuichi Taraki
Summary: This paper presents a computationally efficient method for pose-graph optimization using a multi-resolution representation constructed on a spanning tree. It demonstrates the advantages of the proposed spanning tree-based hierarchy over the previously known serial chain-based hierarchy in terms of sparsity preservation and compatibility with parallel computation. Numerical experiments show that the proposed method outperforms a state-of-the-art solver for large-scale datasets.
IEEE ROBOTICS AND AUTOMATION LETTERS
(2022)
Article
Robotics
Xingyi Li, Han Zhang, Weidong Chen
Summary: This paper presents a 4D radar-based SLAM framework that uses pose graph optimization to achieve accurate and robust pose estimation. The framework filters the raw radar data to reduce noise and estimates ego-velocity to improve registration accuracy. Experimental results demonstrate the precision and robustness of the proposed framework.
IEEE ROBOTICS AND AUTOMATION LETTERS
(2023)
Review
Robotics
Ricardo B. Sousa, Heber M. Sobreira, Antonio Paulo Moreira
Summary: Long-term operation of robots poses challenges to SLAM algorithms, which should adapt to changes while preserving older states. The map size should depend on updating with new information, rather than growing indefinitely based on operation time or trajectory length.
JOURNAL OF FIELD ROBOTICS
(2023)
Article
Optics
Zhiqiang Yan, Hongyuan Wang, Liuchuanjiang Ze, Qianhao Ning, Yinxi Lu
Summary: In this paper, a pose estimation method based on ORBFPFH SLAM is proposed to improve the pose measurement accuracy of the ToF camera for space non-cooperative targets by effectively integrating strength and depth measurement information. The method includes training an ORBFPFH Bag of Words (BoW) model, tracking the pose of the non-cooperative space target based on ORBFPFH feature and optimizing it using the pose graph, and detecting loop closure based on the ORBFPFH BoW model. The proposed method shows improved accuracy in pose estimation and 3D point cloud reconstruction compared to the advanced ORB-SLAM2 algorithm.
Article
Multidisciplinary Sciences
Steven M. Peterson, Satpreet H. Singh, Benjamin Dichter, Michael Scheid, Rajesh P. N. Rao, Bingni W. Brunton
Summary: Understanding the neural basis of human movement in naturalistic scenarios is crucial for advancing neuroscience research. This article introduces a large-scale human neurobehavioral dataset, AJILE12, which includes neural recordings and body pose trajectories, providing data and metadata for exploration and reuse.
Article
Computer Science, Artificial Intelligence
Makoto Okuda, Shinichi Satoh, Yoichi Sato, Yutaka Kidawara
Summary: The paper proposes a method for detecting community structures of graphs by restraining random walk similarity. The method judges starting vertices as being in the same community if the random walkers pass similar sets of vertices. Experimental results demonstrate that the method outperforms previous techniques in terms of accuracy.
IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE
(2021)
Article
Computer Science, Artificial Intelligence
Jia Li, Yongfeng Huang, Heng Chang, Yu Rong
Summary: This paper addresses the problems of node classification and graph classification. It proposes a hierarchical graph modeling approach for the node classification problem, where nodes are graph instances. A novel semi-supervised solution named SEAL-CI is designed to improve accuracy by updating two modules at the graph instance level and the hierarchical graph level.
IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE
(2023)
Article
Computer Science, Artificial Intelligence
Wanfu Gao, Yonghao Li, Liang Hu
Summary: When dealing with high-dimensional multilabel data, we propose a feature selection method that shares latent feature and label structure. By designing an LSS term to share and preserve the latent structure, and employing graph regularization technique to ensure consistency, we achieve better results on multiple evaluation criteria according to experiments.
IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS
(2023)
Article
Automation & Control Systems
Felix Stache, Jonas Westheider, Federico Magistri, Cyrill Stachniss, Marija Popovic
Summary: This paper addresses the problem of adaptive path planning for accurate semantic segmentation using UAVs. An online planning algorithm is proposed to adjust the UAV paths based on detected details in incoming images in order to obtain high-resolution semantic segmentations. The approach is evaluated on real-world data, demonstrating its effectiveness and generality.
ROBOTICS AND AUTONOMOUS SYSTEMS
(2023)
Article
Automation & Control Systems
Hao Dong, Xieyuanli Chen, Simo Sarkka, Cyrill Stachniss
Summary: This paper introduces a novel, accurate, and fast pole extraction approach based on geometric features. The method performs computations directly on range images generated from 3D LiDAR scans and uses the extracted poles as pseudo labels to train a deep neural network for online pole segmentation. Experimental results demonstrate that the proposed method outperforms other approaches in localization on different datasets and LiDAR scanners.
ROBOTICS AND AUTONOMOUS SYSTEMS
(2023)
Article
Automation & Control Systems
Mehul Arora, Louis Wiesmann, Xieyuanli Chen, Cyrill Stachniss
Summary: This paper addresses the problem of building maps of the static aspects of the world by detecting and removing dynamic points. It proposes a method to remove dynamic objects and maintain a high-quality map, and the evaluation results show its superior performance.
ROBOTICS AND AUTONOMOUS SYSTEMS
(2023)
Article
Robotics
Matthias Zeller, Jens Behley, Michael Heidingsfeld, Cyrill Stachniss
Summary: This study proposes a novel approach using radar sensors to perform sparse, single-scan segmentation of moving objects in dynamic environments. The approach includes the Gaussian radar transformer and attentive up and downsampling modules to capture long-range dependencies and achieve superior segmentation quality in diverse environments.
IEEE ROBOTICS AND AUTOMATION LETTERS
(2023)
Article
Robotics
Rodrigo Marcuzzi, Lucas Nunes, Louis Wiesmann, Jens Behley, Cyrill Stachniss
Summary: Autonomous vehicles need to understand their surroundings geometrically and semantically in order to plan and act appropriately in the real world. This paper proposes an approach called MaskPLS to perform panoptic segmentation of LiDAR scans by predicting a set of non-overlapping binary masks and semantic classes, fully avoiding the clustering step.
IEEE ROBOTICS AND AUTOMATION LETTERS
(2023)
Article
Robotics
Nicky Zimmerman, Tiziano Guadagnino, Xieyuanli Chen, Jens Behley, Cyrill Stachniss
Summary: This article presents a method for long-term localization in a changing indoor environment. By utilizing semantic cues and abstract semantic maps, the article proposes a localization framework that combines object detection and camera data with particle filters.
IEEE ROBOTICS AND AUTOMATION LETTERS
(2023)
Article
Robotics
Louis Wiesmann, Lucas Nunes, Jens Behley, Cyrill Stachniss
Summary: This letter focuses on point cloud-based place recognition and proposes a novel neural network architecture to reduce the training time. It extracts local features and computes the similarity between locations based on a global descriptor. By utilizing feature banks, faster training and improved performance are achieved.
IEEE ROBOTICS AND AUTOMATION LETTERS
(2023)
Article
Robotics
Haofei Kuang, Xieyuanli Chen, Tiziano Guadagnino, Nicky Zimmerman, Jens Behley, Cyrill Stachniss
Summary: This letter addresses the problem of estimating a mobile robot's pose in an indoor environment using 2D LiDAR data. It proposes a neural occupancy field method to implicitly represent the scene and synthesizes 2D LiDAR scans for arbitrary robot poses through volume rendering. The synthesized scans are used in an MCL system as an observation model for accurate localization.
IEEE ROBOTICS AND AUTOMATION LETTERS
(2023)
Article
Robotics
Ignacio Vizzo, Tiziano Guadagnino, Benedikt Mersch, Louis Wiesmann, Jens Behley, Cyrill Stachniss
Summary: This article introduces a simple and efficient sensor-based odometry system for accurate pose estimation of a robotic platform. The system utilizes point-to-point ICP matching, adaptive thresholding for correspondence matching, robust kernel, a simple yet widely applicable motion compensation approach, and point cloud subsampling strategy. It can operate under various environmental conditions using different LiDAR sensors.
IEEE ROBOTICS AND AUTOMATION LETTERS
(2023)
Article
Robotics
Jan Weyler, Thomas Laebe, Federico Magistri, Jens Behley, Cyrill Stachniss
Summary: Precision farming robots have the potential to reduce agrochemicals and promote sustainable agriculture. However, a robust plant classification system is needed to identify crops and weeds in diverse agricultural fields.
IEEE ROBOTICS AND AUTOMATION LETTERS
(2023)
Article
Robotics
Louis Wiesmann, Tiziano Guadagnino, Ignacio Vizzo, Nicky Zimmerman, Yue Pan, Haofei Kuang, Jens Behley, Cyrill Stachniss
Summary: This letter introduces a method for mapping the environment using LiDAR point clouds with the goal of obtaining a map representation suitable for robot localization. The researchers utilize a neural network to learn a discretization-free distance field and supervise the network by sampling points along the measured beams. In addition, the paper demonstrates how to perform scan registration and global localization within the neural distance field.
IEEE ROBOTICS AND AUTOMATION LETTERS
(2023)
Article
Robotics
Yue Linn Chong, Jan Weyler, Philipp Lottes, Jens Behley, Cyrill Stachniss
Summary: Agricultural robots have the potential to improve efficiency and sustainability in agriculture. However, machine vision systems used by these robots often perform poorly in new fields or with different robotic platforms. To address this, we propose an approach that can improve the performance without additional manual labeling.
IEEE ROBOTICS AND AUTOMATION LETTERS
(2023)
Article
Robotics
Elias Marks, Matteo Sodano, Federico Magistri, Louis Wiesmann, Dhagash Desai, Rodrigo Marcuzzi, Jens Behley, Cyrill Stachniss
Summary: Measuring plant traits with high throughput allows breeders to select the best cultivars for subsequent breeding generations, improving yield and production of food, feed, and fiber. We use 3D deep learning to build a convolutional neural network that learns to segment individual leaves, automating the breeding process and reducing manual labor. We also propose using an additional neural network to predict leaf quality and discard inaccurate leaf instances.
IEEE ROBOTICS AND AUTOMATION LETTERS
(2023)
Article
Robotics
Benedikt Mersch, Tiziano Guadagnino, Xieyuanli Chen, Ignacio Vizzo, Jens Behley, Cyrill Stachniss
Summary: Mobile robots navigating in unknown environments need to be aware of dynamic objects for mapping, localization, and planning. This letter presents a method that jointly estimates moving objects in the current 3D LiDAR scan and a local map using sparse 4D convolutions. The proposed approach outperforms existing baselines and can be generalized to different types of LiDAR sensors. Results show that the volumetric belief fusion increases the precision and recall of moving object segmentation, even in online mapping scenarios.
IEEE ROBOTICS AND AUTOMATION LETTERS
(2023)
Article
Robotics
Julius Rueckin, Federico Magistri, Cyrill Stachniss, Marija Popovic
Summary: This paper proposes a novel general planning framework for UAVs to autonomously acquire informative aerial images for model retraining. By leveraging multiple acquisition functions and mapping them into probabilistic terrain maps, the framework allows the UAV to adaptively acquire informative images. Experimental results show that the framework maximizes model performance and significantly reduces labeling efforts.
IEEE TRANSACTIONS ON ROBOTICS
(2023)