检索结果-内蒙古大学图书馆

arXiv 2023年

作者： Ausserlechner, Philipp Haberger, David Thalhammer, Stefan Weibel, Jean-Baptiste Vincze, Markus Vision for Robotics Laboratory Automation and Control Institute TU Wien Austria

As robotic systems increasingly encounter complex and unconstrained real-world scenarios, there is a demand to recognize diverse objects. The state-of-the-art 6D object pose estimation methods rely on object-specific training and therefore do not generalize to unseen objects. Recent novel object pose estimation methods are solving this issue using task-specific fine-tuned CNNs for deep template matching. This adaptation for pose estimation still requires expensive data rendering and training procedures. MegaPose for example is trained on a dataset consisting of two million images showing 20,000 different objects to reach such generalization capabilities. To overcome this shortcoming we introduce ZS6D, for zero-shot novel object 6D pose estimation. Visual descriptors, extracted using pre-trained Vision Transformers (ViT), are used for matching rendered templates against query images of objects and for establishing local correspondences. These local correspondences enable deriving geometric correspondences and are used for estimating the object's 6D pose with RANSAC-based PnP. This approach showcases that the image descriptors extracted by pre-trained ViTs are well-suited to achieve a notable improvement over two state-of-the-art novel object 6D pose estimation methods, without the need for task-specific fine-tuning. Experiments are performed on LMO, YCBV, and TLESS. In comparison to one of the two methods we improve the Average Recall on all three datasets and compared to the second method we improve on two datasets. Copyright © 2023, The Authors. All rights reserved.

关键词： Template matching

来源：评论

学校读者我要写书评

暂无评论

Research on optical properties of human hepatoma cells using terahertz time-domain spectroscopy

Research on optical properties of human hepatoma cells using...

引用

2023 International Symposium on Ultrafast Phenomena and Terahertz Waves, ISUPTW 2023

作者： Guan, Hanxiao Qiu, Weihang Ma, Yehao Tian, Liangfei Cao, Yuqi Zhang, Guangxin College of Control Science and Engineering Zhejiang University State Key Laboratory of Industrial Control Technology Hangzhou China Zhejiang University College of Biomedical Engineering and Instrument Science Hangzhou China Ningbo University of Technology Robotics Institute Ningbo China

ISBN: (纸本)9798350317473

Terahertz technology is a non-invasive and label-free inspection method, promising for applications in various fields, especially in biomedical diagnosis. Terahertz radiation is capable for cell physiological and pathological status recognition due to its sensitivity to cell composition and structural information of biomacromolecular. In this paper, the transmission properties of glass, polymethyl methacrylate and polypropylene in THz regime were compared and polypropylene was selected as the petri dish material for cell adhesion due to its high transmission of terahertz wave. The time-domain and frequency-domain spectra, refractive index and absorption coefficient of hepatoma cell line Huh7 and normal cell line MIHA were analyzed. The experimental results showed significant distinction between the terahertz characteristic spectra of hepatoma cells and normal cells. This work can be helpful for applying terahertz technology in cancer detection at the cellular level. © 2023 IEEE.

关键词： Terahertz waves

来源：评论

学校读者我要写书评

暂无评论

A Framework for Automatic Identification of Neural Network Structural Redundancy Based on Reinforcement Learning 5

A Framework for Automatic Identification of Neural Network S...

引用

5th International Conference on Computer Information Science and Artificial Intelligence, CISAI 2022

作者： Wu, Tingting Song, Chunhe Zeng, Peng State Key Laboratory of Robotics Shenyang Institute of Automation Chinese Academy of Sciences Shenyang110016 China Key Laboratory of Networked Control Systems Chinese Academy of Sciences Shenyang110016 China Institutes for Robotics and Intelligent Manufacturing Chinese Academy of Sciences Shenyang110169 China University of Chinese Academy of Sciences Beijing100049 China

ISBN: (纸本)9781510662520

The increasing structure of neural networks makes it difficult to deploy on edge devices with limited computing resources. Network pruning has become one of the most successful model compression methods in recent years. Existing works usually compress models by removing unimportant filters based on the importance. However, the importance-based algorithms tend to ignore the parameters that extract edge features with small criterion values. And recent studies have shown that the existing criteria rely on norm and lead to similar model compression structures. Aiming at the problems of ignoring edge features and manually specifying the pruning rate in current importance-based model pruning algorithms, this paper proposes an automatic recognition framework for neural network structure redundancy based on reinforcement learning. First, we perform cluster analysis on the filters of each layer, and map the filters into a multidimensional space to generate similar sets with different functions. We then propose a criterion for identifying redundant filters within similar sets. Finally, we use reinforcement learning to automatically optimize the cluster dimension, and then determine the optimal pruning rate for each layer to reduce the performance loss caused by pruning. Extensive experiments on various benchmark network architectures and datasets demonstrate the effectiveness of our proposed framework. © 2023 SPIE.

关键词： Reinforcement learning

来源：评论

学校读者我要写书评

暂无评论

Target-driven Object Grasping Method for Mobile Manipulator in Clutter

Target-driven Object Grasping Method for Mobile Manipulator ...

引用

2022 IEEE International Conference on Unmanned Systems, ICUS 2022

作者： Yu, Xinkai Gao, Yongzhuo Dong, Wei Li, Mingyang Liu, Huajian Wang, Chao Harbin Institute of Technology State Key Laboratory of Robotics and System Harbin China Beijing Institute of Control Engineering Beijing China

ISBN: (数字)9781665484565

ISBN: (纸本)9781665484565

The demand for mobile manipulator in the military field is strong, but there is no product that can achieve autonomous grasping in a cluttered environment. In this paper, a target-driven grasp pose detection algorithm is proposed, which can be used to detect the grasp pose of a specified target object in a cluttered environment. After that, based on the lightweight requirements of the military, a mobile manipulator platform is built, and the forward and inverse kinematics model of the mobile manipulator is established. At the same time, a vehicle-arm coordination algorithm is designed to make the grasping efficiency higher. Finally, we use the mobile manipulator to perform autonomous grasping in a cluttered environment to verify the system functionality. © 2022 IEEE.

关键词： Manipulators

来源：评论

学校读者我要写书评

暂无评论

Insulators Defect Detection for Power Grid Based on Improved Yolov7 Algorithm 13

Insulators Defect Detection for Power Grid Based on Improved...

引用

13th International Conference on Intelligent control and Information Processing, ICICIP 2025

作者： Zhang, Zhiwei Zhuang, Yan Liang, Zhida Kong, Shuai Zhu, Shengyi School of Control Science and Engineering Dalian University of Technology Dalian China State Key Laboratory of Robotics Shenyang Institute of Automation Chinese Academy of Sciences Shenyang China Key Laboratory of Special Equipment Safety Testing Technology of Zhejiang Province Zhejiang Academy of Special Equipment Science Hangzhou China

ISBN: (纸本)9798331516147

High-current out-of-phase closed bus is an essential link in the transmission of electric energy and is widely used in thermal power stations, hydropower stations and nuclear power plants. Although these busbars have the advantages of excellent dustproof effect and strong magnetic shielding, they are susceptible to insulator aging and damage in long-term operation, posing a potential safety risk to the power grid. Due to the enclosed space inside the bus and the complexity of external maintenance tasks, manual inspection has problems such as inefficiency and low safety. In order to solve these problems, this paper proposes an improved detection algorithm based on Yolov7 for insulator defect detection in hydropower station scenarios. Because of the particularity of Yolov7 network structure and application scenarios, this paper proposes a Channel Feature Enhancement module to enhance the channel representation ability of the network. In addition, an Edge Feature Enhancement module is introduced to improve the boundary detection of insulators by enhancing the contrast between the target and the background. Considering the limitations of existing studies in specific scenarios and the scarcity of dataset, this paper constructs and accurately annotates an insulator defect detection dataset applied to hydropower station scenarios. Experimental results show that the improved-Yolov7-tiny algorithm in this paper achieves a balance between accuracy and inference speed. © 2025 IEEE.

关键词： Magnetic shielding

来源：评论

学校读者我要写书评

暂无评论

Nonconvex Noise-Tolerant Neural Model for Repetitive Motion of Omnidirectional Mobile Manipulators

引用

IEEE/CAA Journal of Automatica Sinica 2023年第8期10卷 1766-1768页

作者： Zhongbo Sun Shijun Tang Jiliang Zhang Junzhi Yu the Department of Control Engineering Changchun University of TechnologyChangchun 130012China IEEE College of Information Science and Engineering Northeastern UniversityShenyang 110819China the State Key Laboratory for Turbulence and Complex Systems Department of Advanced Manufacturing and RoboticsCollege of EngineeringPeking UniversityBeijing 100871China

Dear Editor,Quadratic programming problems(QPs)receive a lot of attention in various fields of science computing and engineering applications,such as manipulator control[1].Recursive neural network(RNN)is considered t... 详细信息

关键词： directional programming hardware

来源：评论

学校读者我要写书评

暂无评论

A new active rehabilitation training mode for upper limbs based on Tai Chi Pushing Hands

引用

Biomimetic Intelligence & robotics 2024年第3期4卷 82-92页

作者： Xiangpan Li Liaoyuan Li Jianhai Han Bingjing Guo Ganqin Du School of Mechatronics Engineering Henan University of Science and TechnologyLuoyang 471003China Henan Provincial Key Laboratory of Robotics and Intelligent Systems Luoyang 471003China School of Automotive Engineering Hubei University of Automotive TechnologyShiyan 442002China Hubei Key Laboratory of Automotive Power Train and Electronic Control Shiyan 442002China The First Affiliated Hospital of Henan University of Science and Technology Luoyang 471003China

Robot-assisted rehabilitation is a crucial approach to restoring motor function in the ***,the current training trajectory lacks sufficient theoretical or practical support,and the monotony of single-mode training is a *** Chi Pushing Hands,a beneficial and effective daily exercise,has been shown to improve balance function,psychological state,and motor function of the upper extremities in patients recovering from *** address these issues,we propose a new active rehabilitation training that incorporates Tai Chi Pushing Hands movements and yin-yang balance *** training trajectory and direction are encoded by the velocity field and consist of two processes:yang(push)and yin(return).During yang,the limb actively pushes the robot to move,while during yin,the limb actively follows the robot’s *** provide necessary assistance,an admittance controller with self-adaptive parameters is *** addition,we introduce two indexes,the‘Intention Angle’(ϖ)and the time ratio(Γ),to evaluate motion perception *** experiment was conducted on a 4-degree-of-freedom upper limb rehabilitation robot platform,and the subjects were separated into a familiar group and an unfamiliar *** experiment results show that the training could be completed well no matter whether the subject is familiar with Tai Chi Pushing Hands or *** parameters and the movement of the robot can be adjusted based on the interactive force to adapt to the ability of the subject.

关键词： Tai Chi Push Hands Velocity field Self-adaptive admittance Upper limb rehabilitation Active training

来源：评论

学校读者我要写书评

暂无评论

An Improved DeeplabV3+ Tile Image Positioning Method

An Improved DeeplabV3+ Tile Image Positioning Method

引用

2024 International Conference Optoelectronic Information and Optical Engineering, OIOE 2024

作者： Li, Xiaoyang Wang, Rui Tong, Xing Niu, Lianqiang Shenyang University of Technology Shenyang China Key Laboratory of Network Control Systems Chinese Academy of Sciences China Shenyang Institute of Automation Chinese Academy of Sciences China Institutes for Robotics and Intelligent Manufacturing Chinese Academy of Sciences Shenyang China

ISBN: (数字)9781510688209

ISBN: (纸本)9781510688193

In order to improve the detection accuracy and efficiency of surface defects on ceramic tiles, a detection model based on DeepLabV3+and combined with MobileNetv2 for feature extraction was constructed. Using ECA attention mechanism and multi branch receptive field cascade fusion structure to expand the sampling range of global and local element features for the empty space pyramid pooling module ASPP of the original model. Meanwhile, the model decoder uses depthwise separable convolutions instead of the original standard convolutions. Experimental results have shown that the improved model has an MPA value of 96.12%, a MIou value of 97.29%, an F1 value of 90.03%, and an Accuracy value of 99.89%. Compared with the previous model, the improved model's indicators have increased by 4.07%, 4.44%, 2.04%, and 1.02%, respectively. The segmentation accuracy and detection speed have been significantly improved. The experimental results fully demonstrate that the algorithm proposed in this article can effectively complete the task of detecting surface defects on ceramic tiles. © 2025 SPIE.

关键词： Image segmentation

来源：评论

学校读者我要写书评

暂无评论

Lidar-Inertial and Downward Vision Based Mobile Robot SLAM in Digital Workshops 13

Lidar-Inertial and Downward Vision Based Mobile Robot SLAM i...

引用

13th IEEE Data Driven control and Learning Systems Conference, DDCLS 2024

作者： Zhang, Yanchao Geng, Hao Liu, Bin Zhang, Yinlong Yuan, Xudong Zhang, Sichao School of Information Science and Engineering Shenyang University of Technology Shenyang110020 China Guangzhou Institute of Industrial Intelligence Guangzhou511458 China State Key Laboratory of Robotics Shenyang Institute of Automation Chinese Academy of Sciences Shenyang110016 China Key Laboratory of Networked Control Systems Chinese Academy of Sciences Shenyang110016 China Institutes for Robotics and Intelligent Manufacturing Chinese Academy of Sciences Shenyang110169 China

ISBN: (纸本)9798350361674

In order to solve the problems of long-term cumulative error and the absence of absolute pose in the existing Lidar-Inertial Measurement Unit (IMU) fusion methods for Automated Guided Vehicle (AGV) Simultaneous Localization and Mapping (SLAM), we propose a new SLAM method for mobile robot, integrating Lidar, IMU, and a downward vision camera. Firstly, we tightly couple Lidar-IMU and downward vision camera using a multimodal fusion framework. Downward visual measurements are integrated as QR factors into the factor graph, providing absolute pose information to correct cumulative error in the Lidar-IMU system and enable the joint pose estimation of the AGV. Secondly, an enhanced Iterative Closest Point (ICP) point cloud matching algorithm has been developed to correct the incorrect pose. This approach achieves Scan-to-Map alignment by employing a coarse-to-fine matching strategy, thereby enhancing the accuracy of point cloud alignment. Finally, the extended Kalman filter is used to optimize the Lidar pose drift, as well as the Lidar mapping effectively by using the Lidar-IMU as the measurements and the downward visual measurements as the updates. The algorithm has been extensively evaluated on the developed AGV experimental platform. The experimental results show that the position error is less than 0.02 meters and the orientation error is less than 1°, which solves the problems of long-term error accumulation and the lack of absolute pose information for AGV navigation. © 2024 IEEE.

关键词： Extended Kalman filters

来源：评论

学校读者我要写书评

暂无评论

DEGAS - Deformation Estimation based on Magnetic Gradient Tensor Contraction for Soft Robot Shape Servoing

DEGAS - Deformation Estimation based on Magnetic Gradient Te...

引用

IEEE International Conference on Soft robotics (RoboSoft)

作者： Jelena Vuletić Bruno Marić Matko Orsag LARICS Laboratory for Robotics and Intelligent Control Systems University of Zagreb Faculty of Electrical Engineering and Computing Zagreb Croatia

ISBN: (数字)9798331520205

ISBN: (纸本)9798331520212

This work proposes the first derivation, implementation, and experimental validation of magnetic-based proprioceptive sensing method for soft robotic applications. In our proposed approach, the magnetic sensing system measures gradient tensor contractions that can be directly related to the shape of a deformable plastering tool. Custom-designed and 3D-printed plastering tool embeds two identical permanent magnets that generate a non-symmetric magnetic field tracked by the proposed sensor. Seamless real-time control is enabled with the sensor sampling rate of 1 kHz. Classical linear control is synthesized for the scraper bending angle and orientation control. The tool, sensor, and the proposed control system are validated on robotic plastering and painting tasks.

关键词： Magnetic field measurement Tensors Deformation Shape Propioception Estimation Bending Soft robotics Sensors Painting

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：