检索结果-内蒙古大学图书馆

Conference on Computer Vision and Pattern Recognition (CVPR)

作者： Masashi Yokozuka Shuji Oishi Simon Thompson Atsuhiko Banno Robot Innovation Research Center National Institute of Advanced Industrial Science and Technology (AIST) Japan National Institute of Advanced Industrial Science and Technology Tsukuba Japan

In this paper, we propose a novel indirect monocular simultaneous localization and mapping (SLAM) algorithm called "VITAMIN-E," which is highly accurate and robust as a result of tracking extremely dense feature points. Typical indirect methods have difficulty in reconstructing dense geometry because of their careful feature point selection for accurate matching. Unlike conventional methods, the proposed method processes an enormous number of feature points using the tracking local extrema of curvature based on dominant flow estimation. Because this may lead to high computational cost during bundle adjustment, we propose a novel optimization technique called the "subspace Newton's method" that significantly improves the computational efficiency of bundle adjustment by partially updating the variables. We concurrently generate meshes from the reconstructed points and merge them for an entire three-dimensional(3D) model. Experimental results on the SLAM benchmark EuRoC demonstrated that the proposed method outperformed state-of-the-art SLAM methods such as DSO, ORB-SLAM, and LSD-SLAM, both in terms of accuracy and robustness in trajectory estimation. The proposed method simultaneously generated significantly detailed 3D geometry as a result of the dense feature points in real time using only a CPU.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Spurious reconstruction from brain activity

引用

Neural Networks 2025年 190卷

作者： Ken Shirakawa Yoshihiro Nagano Misato Tanaka Shuntaro C. Aoki Yusuke Muraki Kei Majima Yukiyasu Kamitani Graduate School of Informatics Kyoto University Sakyo-ku Kyoto 606-8501 Japan Computational Neuroscience Laboratories Advanced Telecommunications Research Institute International Seika Soraku Kyoto 619-0288 Japan Institute for Quantum Life Science National Institutes for Quantum Science and Technology Anagawa Chiba 263-8555 Japan Guardian Robot Project RIKEN Seika Soraku Kyoto 619-0288 Japan

Advances in brain decoding, particularly in visual image reconstruction, have sparked discussions about the societal implications and ethical considerations of neurotechnology. As reconstruction methods aim to recover visual experiences from brain activity and achieve prediction beyond training samples (zero-shot prediction), it is crucial to assess their capabilities and limitations to inform public expectations and regulations. Our case study of recent text-guided reconstruction methods, which leverage a large-scale dataset (Natural Scenes Dataset, NSD) and text-to-image diffusion models, reveals critical limitations in their generalizability, demonstrated by poor reconstructions on a different dataset. UMAP visualization of the text features from NSD images shows limited diversity with overlapping semantic and visual clusters between training and test sets. We identify that clustered training samples can lead to “output dimension collapse,” restricting predictable output feature dimensions. While diverse training data improves generalization over the entire feature space without requiring exponential scaling, text features alone prove insufficient for mapping to the visual space. Our findings suggest that the apparent realism in current text-guided reconstructions stems from a combination of classification into trained categories and inauthentic image generation (hallucination) through diffusion models, rather than genuine visual reconstruction. We argue that careful selection of datasets and target features, coupled with rigorous evaluation methods, is essential for achieving authentic visual image reconstruction. These insights underscore the importance of grounding interdisciplinary discussions in a thorough understanding of the technology’s current capabilities and limitations to ensure responsible development.

关键词： Brain decoding Visual image reconstruction Naturalistic approach NeuroAI

来源：评论

学校读者我要写书评

暂无评论

A novel hierarchical control strategy for biped robot walking on uneven terrain

A novel hierarchical control strategy for biped robot walkin...

引用

IEEE-RAS International Conference on Humanoid robots

作者： Chencheng Dong Xuechao Chen Zhangguo Yu Zelin Huang Qingqing Li Qinqin Zhou Qiang Huang School of Mechantronical Engineering Beijing Institute of Technology Beijing China Bionic robot mechanism perception and control laboratory Beijing China Beijing Advanced Innovation Center for Intelligent Robotics and Systems Beijing Institute of Technology Beijing China

ISBN: (数字)9781538676301

ISBN: (纸本)9781538676318

When position-controlled biped robot is blind walking on a uneven terrain at a high speed, huge foot contact impacts will be generated. However, traditional admitance control can't absorb the impact and stabilize the robot due to its slow response and Incompleteness. In this paper, we propose a control strategy including respectively designed swing leg control and support leg control with a new approach of control transition. For Swing leg control, double spring damping model is presented to optimize the admitance controller with faster response and better robustness, and a active foot height controller is also proposed to reduce the impact further. On the other hand, the control transition is accomplished by using a bionic fuzzy control. As a result, the foot contact impact can be reduced and the robot can blind walk fast on uneven terrain. Finally, the validity of the proposed strategy is confirmed by the simulation.

关键词： Legged locomotion Foot Trajectory Springs Adaptation models robot kinematics

来源：评论

学校读者我要写书评

暂无评论

advanced SURF Features Based Flexible Object Detection

Advanced SURF Features Based Flexible Object Detection

引用

IEEE International Conference on robotics and Biomimetics

作者： Di Lv Yunfu Deng Zhihao Li Qujiang Lei Bo Liang Jie Xu Xiuhao Li intelligent robot & equipment center Guangzhou Institute of Advanced Technology Chinese Academy of Sciences Guangzhou China

On the modern assembly line, the posture of flexible products can be distorted or stacked on each other, which is difficult for methods currently in effect to ensure the precision with current methods. To solve this problem, we demonstrate our method on ABB Yumi robot for flexible object detection based on SURF features with color channel prior and marginal feature optimization. Our method achieved 97.99% accuracy, which is 7.67% better than vanilla SURF.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Research on Trajectory Planning of 6-DOF Cutting-robot in Machining Complex Surface 2

Research on Trajectory Planning of 6-DOF Cutting-robot in Ma...

引用

2nd International Conference on Mechanical, System and Control Engineering, ICMSC 2018

作者： Jiao, Sheng-Xi Wang, Hao Xia, Lin-Lin Zhang, Shuai University of Northeast Electric Power in Jilin Advanced Sensor Technology Institute Jilin China University of Northeast Electric Power in Jilin Research Institute of Robot and Image Visual Detection Technology Jilin China

It is important and difficult for the complicated surface processing in mechanical industry. In this paper, an improved algorithm for trajectory planning is proposed in impeller surface processing by using 6-DOF cutting-robot. Taking a single finished path of the impeller blade as an example, the feedrate of the cutter, bow height error, cutter-orientation and position are planned by the B-spline interpolation algorithm, the best cutting trajectory is obtained. On the basis of trajectory planning, the optimal movement scheme of 6-DOF cutting-robot joints is obtained, the 6-DOF cutting-robot feedrate and trajectory smooth transition is achieved and the joints movement adaptive adjustment is completed. Finally, the angles, the angular velocitys of the joints and their interrelated properties are analyzed. The research works indicate that the robot joint angle curves are continuous and stable, which has met the requirements of smooth movement of the robot, and the results show that the trajectory planning is effective and practical. © The Authors, published by EDP Sciences, 2018.

关键词： Interpolation

来源：评论

学校读者我要写书评

暂无评论

Dynamic Modeling Method of Spatial Passive Over-constrained Parallel Mechanism Based on Newton Euler Method

引用

Jixie Gongcheng Xuebao/Journal of Mechanical Engineering 2020年第11期56卷 48-57页

作者： Li, Yongquan Guo, Yu Zhang, Yang Zhang, Lijie Parallel Robot and Mechatronic System Laboratory of Hebei Province Yanshan University Qinhuangdao066004 China Key Laboratory of Advanced Forging & Stamping Technology and Science Yanshan University Ministry of Education of China Qinhuangdao066004 China Hebei Provincial Key Laboratory of Heavy Machinery Fluid Power Transmission and Control Yanshan University Qinhuangdao066004 China

The 4-RPTR four-degree-of-freedom space passive over-constrained (redundant constrained) parallel mechanism is taken as the research object, and the Newton Euler method combined with the complementary deformation coordination equation is used to analyze the dynamics. The number of dynamic equations based on Newton's Euler method is 78, but due to the existence of passive over-constraints of the mechanism, the unknown variables of the required solutions are 80. When the eight deformation coordination equations are added according to the method of planar over-constrained mechanism, it will be linearly related to the 78 equations built, resulting in the equations being unable to solve. Based on this, when proposing the dynamics modeling of passive over-constrained parallel mechanism, the reason for the passive over-constraint of the mechanism should be analyzed firstly. On this basis, adding the deformation of the member caused by the redundant constrained force (torque) and the position and posture error of the terminal platform as a complementary equation ensures that the equations are linearly independent and can be solved. To this end, the relationship between the deformation of the rod and the position and posture error of the terminal platform caused by the rotational joint restraining torque directly causing the redundant constraint is introduced, and eight deformation coordination equations are added, and six unknown variables are introduced, thereby obtaining 86 equations. Containing 86 unknowns, and the equations are linearly independent, the dynamic solution model of the mechanism is obtained. Based on matlab and adams software, the dynamics simulation of the mechanism is carried out. When the same motion trajectory is given at the terminal of the platform, the correctness of the dynamic model can be seen by comparing the driving force variation curves of the two methods. It provides a new idea for dynamic modeling of this type of space passive over-constrai

关键词： Dynamics

来源：评论

学校读者我要写书评

暂无评论

Design and study of a biomimetic wall-climbing robot with spine wheels and a vacuum suction cup

Design and study of a biomimetic wall-climbing robot with sp...

引用

IEEE International Conference on Real-time Computing and robotics (RCAR)

作者： Jinfu Liu Linsen Xu Shouqi Chen Hong Xu Gaoxin Cheng Lei Liu Jiajun Xu Hefei Institutes of Physical Science CAS and University of Science and Technology of China Hefei China Hefei Institutes of Physical Science CAS and the Key Laboratory of Biomimetic Sensing and Advanced Robot Technology Changzhou Anhui Province China University of Science and Technology of China Hefei China

ISBN: (数字)9781728137261

ISBN: (纸本)9781728137278

Wall-climbing robots can be widely applied in search and rescue, nuclear power plant maintenance, pressure pipeline inspection and so on, as a kind of special robots in extreme environment. However, the traditional single motion mode can not ensure the stable adhesion of the robot on complex wall surfaces. Inspired by the structure characteristics of flies and clingfish, a biomimetic flexible spine wheel and an eddy suction cup mechanism are proposed. Then, based on the law of mechanism configuration synthesis, a novel wall-climbing locomotion mechanism is developed by equipping these above biomimetic mechanisms. With the elastic mechanics theory of finite element, the original stiffness matrix of a single spine wheel at different stiffness is deduced. Based on the transformation mechanism of rotation matrix and the original stiffness matrix of the spine wheel, an innovative contact model is set up to analyze the forces acting on a single spine wheel. Beyond that, mechanical analysis model between suction force and load is also presented, and the simulation results show that the mechanical model is effective and reasonable in the selection of vacuum component. Finally, we fabricated the wall-climbing robot adopting the technology of 3D printing and precision machining to verify its climbing performance. Meanwhile, experiment results present the robot could climb stably on the 0-360° wall surfaces.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Decision-Level Fusion Method for Emotion Recognition using Multimodal Emotion Recognition Information 15

Decision-Level Fusion Method for Emotion Recognition using M...

引用

15th International Conference on Ubiquitous robots, UR 2018

作者： Song, Kyu-Seob Nho, Young-Hoon Seo, Ju-Hwan Kwon, Dong-Soo Department of Robotics Program Korea Advanced Institute of Science and Technology Human-Robot Interaction Research Center Korea Republic of Department of Mechanical Engineering Korea Advanced Institute of Science and Technology Human-Robot Interaction Research Center Korea Republic of

ISBN: (纸本)9781538663349

Human emotion recognition is an important factor for social robots. In previous research, emotion recognizers with many modalities have been studied, but there are several problems that make recognition rates lower when a recognizer is applied to a robot. This paper proposes a decision level fusion method that takes the outputs of each recognizer as an input and confirms which combination of features achieves the highest accuracy. We used EdNet, which was developed in KAIST based Convolutional Neural Networks (CNNs), as a facial expression recognizer and a speech analytics engine developed for speech emotion recognition. Finally, we confirmed a higher accuracy 43.40% using an artificial neural network (ANN) or the k-Nearest Neighbor (k-NN) algorithm for classification of combinations of features from EdN et and the speech analytics engine. © 2018 IEEE.

关键词： Engines

来源：评论

学校读者我要写书评

暂无评论

Effect of the configuration and concentration of B-doping on the thermal properties and structural stability of graphene: A theoretical study

引用

Diamond and Related Materials 2025年 156卷

作者： Li, Dongbo Da, Liangguo Huang, Kaifeng Li, Dong Yu, Liang Wang, Haini Yu, Yaqi Wu, Long Tang, Yunqing School of Mechanical and Electrical Engineering Huainan Normal University Huainan232038 China Human-computer collaborative robot Joint Laboratory of Anhui Province Huainan232038 China School of Mechanical Engineering Shandong University Jinan250061 China State Key Laboratory of Advanced Equipment and Technology for Metal Forming Shandong University Jinan250061 China Key Laboratory of High Efficiency and Clean Mechanical Manufacture Ministry of Education Shandong University Jinan250061 China

The thermal properties and structural stability of B-doped graphene were investigated by employing density functional theory (DFT), ab initio molecular dynamics (AIMD), and molecular dynamics (MD) simulations, with a focus on the effects of doping configuration and concentration. DFT calculations indicate that B-doping reduces the cohesive energy of graphene, consequently diminishing its structural stability. Among the three B double-doping configurations, only the para configuration exhibits a phonon dispersion without imaginary frequencies, whereas both meta and ortho configurations show imaginary frequencies, indicating structural instability. The AIMD results further demonstrate that B-doped graphene preferentially undergoes cleavage at B[sbnd]B bonds, leading to decreased structural stability. MD results show that B-doping significantly reduces the thermal conductivity of graphene. At a doping concentration is 0.6 %, the thermal conductivity of meta and para configurations decreases to 48.08 % and 54.6 % of that of the pristine graphene, respectively. The results of phonon properties show that B-doping reduces the phonon participation rate and induces strong phonon localization at the B-doped site. These results and conclusions will provide valuable theoretical insights for the thermal design of graphene-based semiconductor devices. © 2025 Elsevier B.V.

关键词： Temperature distribution

来源：评论

学校读者我要写书评

暂无评论

Higher-Order Kinematics Modeling of 3-RRS Parallel Mechanism Based on CGA

Research Square

引用

Research Square 2020年

作者： Wang, Chang Zhao, Tie-Shi Li, Er-Wei Zhao, Yan-Zhi Bian, Hui Geng, Ming-Chao Parallel Robot and Mechatronic System Laboratory Yanshan University Qinhuangdao066004 China Key Laboratory of Advanced Forging & Stamping Technology and Science of Ministry of National Education Yanshan University Qinhuangdao066004 China School of Mechanical Engineering Hebei University of Architecture Zhangjiakou075000 China

Higher-order kinematics of mechanisms has been applied in servo motor control, human-robot interaction and machinery life design fields, etc. The representations of acceleration and jerk by screws have been fully developed by researchers with the methods of the differential of the matrix representation of SE(3) group. Clifford algebra, which is tighter and with higher computational efficiency than the matrix method, is another representation of the motions of rigid bodies. It has been used in position kinematics, grub task motion planning, and robot vision for its convenience of geometric representations and calculations. As far as we know, the work of higher-order kinematics of mechanisms based on Clifford algebra is rare. First, after recalling the based theory of motion representation in conformal geometric algebra (CGA), the mathematical relationships between flag and motor are built. Second, a method for the higher-order kinematics modeling of serial chain mechanisms is proposed. Finally, the higher-order kinematics of the 3-RRS parallel mechanism is built to prove the correctness of the algorithm. This work further enriches the application of CGA for the higher-order kinematics modeling of parallel mechanisms. © 2020, CC BY.

关键词： Kinematics

来源：评论

学校读者我要写书评

暂无评论

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案：

请选择收藏分类：

通借通还

建议与咨询 留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案： 新增检索档案 确定 取消

请选择收藏分类： 新增自定义分类 确定 取消

通借通还

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

请选择保存的检索档案：

请选择收藏分类：