检索结果-内蒙古大学图书馆

arXiv 2024年

作者： Wang, Ze Li, Yang Xu, Long Shi, Hao Ma, Zunwang Chu, Zhen Li, Chao Gao, Fei Yang, Kailun Wang, Kaiwei State Key Laboratory of Modern Optical Instrumentation Zhejiang University China School of Robotics and National Engineering Research Center of Robot Visual Perception and Control Technology Hunan University China State Key Laboratory of Industrial Control Technology Zhejiang University China DeepRobotics Co. Ltd. China

Dynamic jumping on high platforms and over gaps differentiates legged robots from wheeled counterparts. Compared to walking on rough terrains, dynamic locomotion on abrupt surfaces requires fusing proprioceptive and exteroceptive perception for explosive movements. In this paper, we propose SF-TIM (Simple Framework combining Terrain Imagination and Measurement), a single-policy method that enhances quadrupedal robot jumping agility, while preserving their fundamental blind walking capabilities. In addition, we introduce a terrain-guided reward design specifically to assist quadrupedal robots in high jumping, improving their performance in this task. To narrow the simulation-to-reality gap in quadrupedal robot learning, we introduce a stable and high-speed elevation map generation framework, enabling zero-shot simulation-to-reality transfer of locomotion ability. Our algorithm has been deployed and validated on both the small-/large-size quadrupedal robots, demonstrating its effectiveness in real-world applications: the robot has successfully traversed various high platforms and gaps, showing the robustness of our proposed approach. A demo video has been made available at https://***/SF-TIM. Copyright © 2024, The Authors. All rights reserved.

关键词： Multipurpose robots

来源：评论

学校读者我要写书评

暂无评论

Posture-based Leader-follower Formation for Multiple Mobile robot Systems Using Data-based Adaptive Iterative Learning control

Posture-based Leader-follower Formation for Multiple Mobile ...

引用

Chinese Automation Congress (CAC)

作者： Zhuolei Chaochen Yaonan Wang Haoran Tan Jiaming Zhou Qiong Lin College of Electrical and Information Engineering Hunan University National Engineering Research Center for Robot Visual Perception and Control Changsha China

Multiple mobile robot systems (MMRSs) have great potential in various fields and their reliable application depends on accurate formation among robots. Focusing on the formation problem of MMRSs, a posture-based adaptive iterative learning control scheme (PB-AILC) is proposed in this paper. Firstly, the leader-follower formation model is built. Then PB-AILC scheme is introduced to optimize the control linear and angular velocity of follower robots at each iteration based on the robots’ global posture only, guaranteeing that followers can steadily track their leader with prescribed relative distance and orientation. Meanwhile, we analyze its feasibility mathematically. Finally, numerical simulations illustrate the validity of the PB-AILC scheme and we compare PB-AILC with an iterative learning control (ILC) method to demonstrate its superiority.

关键词： Adaptation models Adaptive systems Uncertainty Process control Numerical simulation Real-time systems Numerical models

来源：评论

学校读者我要写书评

暂无评论

Semi-supervised Cloud Edge Collaborative Power Transmission Line Insulator Anomaly Detection Framework 11th

Semi-supervised Cloud Edge Collaborative Power Transmission ...

引用

11th International Conference on Image and Graphics, ICIG 2021

作者： Yang, Yanqing Mao, Jianxu Zhang, Hui Chen, Yurong Zhong, Hang Huang, Zhihong Wang, Yaonan National Engineering Laboratory of Robot Visual Perception and Control Technology Hunan University Hunan Changsha China State Grid Hunan Electric Power Corporation Limited Research Institute Hunan Changsha China

ISBN: (纸本)9783030873547

The widely deployed power transmission line expedites developing the age of electricity. Thus, it is necessary to maintain a power system with a great quantity of manpower and material resources, especially for crucial equipment, such as insulator string. However, the current main inspection method relies on artificial with the problem of time-consuming and labor-intensive. There is a trend of utilizing deep learning techniques on unmanned aerial vehicles (UAVs) to accomplish the inspection task, but its development is restricted by the limitation of energy. In this paper, we propose a semi-supervised cloud edge collaborative insulator string anomaly detection framework. Specifically, an anchor-free object detector is deployed on the edge device for locating the insulator. On the cloud side, we propose a generative insulator defect detection model based on the autoencoder (AE) with a generator-discriminator pattern. Particularly, we introduce the variational memory encoder-decoder architecture to model defect-free insulator data distribution. Furthermore, the adversarial strategy is employed to regularize the generated data space with input data space. In the end, the anomaly can be detected if its data space is an outlier of training defect-free distribution. Comprehensive experiments demonstrate that our method can effectively reduce the computational load, meanwhile archiving superior performance, including accuracy (0.968) and recall (0.985), for defect recognition using a standard insulator data set. © 2021, Springer Nature Switzerland AG.

关键词： Deep learning

来源：评论

学校读者我要写书评

暂无评论

Unsupervised Deep Homography Estimation based on Transformer

Unsupervised Deep Homography Estimation based on Transformer

引用

International Conference on Advanced robotics and Mechatronics (ICARM)

作者： Tianjian Jiang Qiu Fang Qing Zhu Yaonan Wang Zhen Zhou Lin Chen Jiaming Zhou Yuefan Luo Chengzhong Wu National Engineering Research Center for Robot Visual Perception and Control Technology Hunan University Changsha Hunan China Jiangxi Communication Terminal Industry Technology Research Institute.

Homography estimation is a crucial problem in computer vision, which aims to provide an optimal transformation matrix for aligning images captured from different viewpoints. Current methods extract shallow features from image pairs and introduce learnable mask modules to improve homography estimation performance. However, they struggle to capture long-term dependencies between features and comprehend the global structures of image features. A deep unsupervised homography learning framework is proposed in this paper, consisting of a weight-sharing feature extraction network and a homography estimation network based on the Transformer model. The former extracts the local features of images, while the latter learns the correlation between them and understands the global features of images, enabling the algorithm to better estimate the homography of unaligned images. Experimental results demonstrate that the proposed method outperforms the advanced methods for estimating homography matrices in the CA-Unsupervised dataset.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Accurate RGB-D SLAM in dynamic environments based on dynamic visual feature removal

引用

Science China(Information Sciences) 2022年第10期65卷 256-269页

作者： Chenxin LIU Jiahu QIN Shuai WANG Lei YU Yaonan WANG Department of Automation University of Science and Technology of China Institute of Artificial Intelligence Hefei Comprehensive National Science Center College of Electrical and Information Engineering Hunan University National Engineering Research Center of Robot Visual Perception and Control Technology Hunan University

visual localization is considered an essential capability in robotics and has attracted increasing interest for the past few years. However, most proposed visual localization systems assume that the surrounding environment is static, which is difficult to maintain in real-world scenarios due to the presence of moving objects. In this paper, we present DFR-SLAM, a real-time and accurate RGB-D SLAM based on ORB-SLAM2 that achieves satisfactory performance in a variety of challenging dynamic scenarios. At the core of our system lies a motion consensus filtering algorithm estimating the initial camera pose and a graph-cut optimization framework combining long-term observations, prior information, and spatial coherence to jointly distinguish dynamic and static visual features. Other systems for dynamic environments detect dynamic components by using the information from short time-span frames, whereas our system uses observations from a long period of keyframes. We evaluate our system using dynamic sequences from the public TUM dataset, and the evaluation demonstrates that the proposed system outperforms the original ORB-SLAM2 system significantly. In addition, our system provides competitive localization accuracy with satisfactory real-time performance compared to closely related SLAM systems designed to adapt to dynamic environments.

关键词： SLAM dynamic environments indoor localization graph-cut robot navigation

来源：评论

学校读者我要写书评

暂无评论

Steady Tracker: Tracking a Target Stably Using a Quadrotor

Steady Tracker: Tracking a Target Stably Using a Quadrotor

引用

2022 IEEE International Conference on robotics and Biomimetics, ROBIO 2022

作者： Li, Hongwen Zhong, Hang Lv, Yongsheng Sha, Jianjun Long, Yu Wang, Yaonan Harbin Engineering University Harbin150001 China Qingdao Innovation and Development Center Harbin Engineering University Qingdao266000 China School of Robotics National Engineering Research Center for Robot Visual Perception and Control Hunan University China College of Electrical and Information Engineering Hunan University Changsha410082 China

ISBN: (纸本)9781665481090

Maneuvering target tracking of Unmanned Aerial Vehicle(UAV) in cluttered environments is a challenging issue owing to the unknown motion intention of the target and the complex moving environments. As the complexity of the environment increases, stable and secure target tracking is increasingly difficult to guarantee. To address the issue, this paper proposes a stable quadrotor tracking solution. The proposed solution contains two parts: target motion prediction and tracking path searching. The target motion prediction method predicts the future target motion based on the obtained target observations while considering observation noise and prediction errors. The tracking path searching method utilizes a sampling-based search method, using the homotopy of paths to ensure that the tracking path and the target position are in the same space. Finally, simulations, real-world experiments and statistical analysis verify the correctness and effectiveness of the proposed approach. © 2022 IEEE.

关键词： Antennas

来源：评论

学校读者我要写书评

暂无评论

Materobot: Material Recognition in Wearable robotics for People with visual Impairments

MateRobot: Material Recognition in Wearable Robotics for Peo...

引用

IEEE International Conference on robotics and Automation (ICRA)

作者： Junwei Zheng Jiaming Zhang Kailun Yang Kunyu Peng Rainer Stiefelhagen Institute for Robotics and Anthropomatics Karlsruhe Institute of Technology Karlsruhe Germany Department of Engineering Science University of Oxford UK School of Robotics and the National Engineering Research Center of Robot Visual Perception and Control Technology Hunan University Changsha China

ISBN: (数字)9798350384574

ISBN: (纸本)9798350384581

People with visual Impairments (PVI) typically recognize objects through haptic perception. Knowing objects and materials before touching is desired by the target users but under-explored in the field of human-centered robotics. To fill this gap, in this work, a wearable vision-based robotic system, MATErobot, is established for PVI to recognize materials and object categories beforehand. To address the computational constraints of mobile platforms, we propose a lightweight yet accurate model MATEViT to perform pixel-wise semantic segmentation, simultaneously recognizing both objects and materials. Our methods achieve respective 40.2% and 51.1% of mIoU on COCOStuff-10K and DMS datasets, surpassing the previous method with +5.7% and +7.0% gains. Moreover, on the field test with participants, our wearable system reaches a score of 28 in the NASA-Task Load Index, indicating low cognitive demands and ease of use. Our MATErobot demonstrates the feasibility of recognizing material property through visual cues and offers a promising step towards improving the functionality of wearable robots for PVI. The source code has been made publicly available at MATErobot.

关键词： visualization Source coding Semantic segmentation visual impairment Semantics Wearable robots Mobile applications

来源：评论

学校读者我要写书评

暂无评论

Towards Precise 3D Human Pose Estimation with Multi-Perspective Spatial-Temporal Relational Transformers

arXiv

引用

arXiv 2024年

作者： Jiao, Jianbin Cheng, Xina Chen, Weijie Yin, Xiaoting Shi, Hao Yang, Kailun School of Artificial Intelligence Xidian University China State Key Laboratory of Extreme Photonics and Instrumentation Zhejiang University China School of Robotics Hunan University China National Engineering Research Center of Robot Visual Perception and Control Technology Hunan University China

3D human pose estimation captures the human joint points in three-dimensional space while keeping the depth information and physical structure. That is essential for applications that require precise pose information, such as human-computer interaction, scene understanding, and rehabilitation training. Due to the challenges in data collection, mainstream datasets of 3D human pose estimation are primarily composed of multi-view video data collected in laboratory environments, which contains rich spatial-temporal correlation information besides the image frame content. Given the remarkable self-attention mechanism of transformers, capable of capturing the spatial-temporal correlation from multi-view video datasets, we propose a multi-stage framework for 3D sequence-to-sequence (seq2seq) human pose detection. Firstly, the spatial module represents the human pose feature by intra-image content, while the frame-image relation module extracts temporal relationships and 3D spatial positional relationship features between the multi-perspective images. Secondly, the self-attention mechanism is adopted to eliminate the interference from non-human body parts and reduce computing resources. Our method is evaluated on Human3.6M, a popular 3D human pose detection dataset. Experimental results demonstrate that our approach achieves state-of-the-art performance on this dataset. The source code will be available at https://***/WUJINHUAN/3D-human-pose. Copyright © 2024, The Authors. All rights reserved.

关键词： Human computer interaction

来源：评论

学校读者我要写书评

暂无评论

Towards Single-Lens controllable Depth-of-Field Imaging via Depth-Aware Point Spread Functions

arXiv

引用

arXiv 2024年

作者： Qian, Xiaolong Jiang, Qi Gao, Yao Gao, Shaohua Yi, Zhonghua Sun, Lei Wei, Kai Li, Haifeng Yang, Kailun Wang, Kaiwei Bai, Jian State Key Laboratory of Extreme Photonics and Instrumentation Zhejiang University Hangzhou310027 China School of Robotics Hunan University Changsha410012 China National Engineering Research Center of Robot Visual Perception and Control Technology Hunan University Changsha410082 China

controllable Depth-of-Field (DoF) imaging commonly produces amazing visual effects based on heavy and expensive high-end lenses. However, confronted with the increasing demand for mobile scenarios, it is desirable to achieve a lightweight solution with Minimalist Optical Systems (MOS). This work centers around two major limitations of MOS, i.e., the severe optical aberrations and uncontrollable DoF, for achieving single-lens controllable DoF imaging via computational methods. A Depth-aware controllable DoF Imaging (DCDI) framework is proposed equipped with All-in-Focus (AiF) aberration correction and monocular depth estimation, where the recovered image and corresponding depth map are utilized to produce imaging results under diverse DoFs of any high-end lens via patch-wise convolution. To address the depth-varying optical degradation, we introduce a Depth-aware Degradation-adaptive Training (DA2T) scheme. At the dataset level, a Depth-aware Aberration MOS (DAMOS) dataset is established based on the simulation of Point Spread Functions (PSFs) under different object distances. Additionally, we design two plug-and-play depth-aware mechanisms to embed depth information into the aberration image recovery for better tackling depth-aware degradation. Furthermore, we propose a storage-efficient Omni-Lens-Field model to represent the 4D PSF library of various lenses. With the predicted depth map, recovered image, and depth-aware PSF map inferred by Omni-Lens-Field, single-lens controllable DoF imaging is achieved. To the best of our knowledge, we are the first to explore the single-lens controllable DoF imaging solution. Comprehensive experimental results demonstrate that the proposed framework enhances the recovery performance, and attains impressive single-lens controllable DoF imaging results, providing a seminal baseline for this field. The source code and the established dataset will be publicly available at https://***/XiaolongQian/DCDI. Copyright © 2024, The A

关键词： Optical transfer function

来源：评论

学校读者我要写书评

暂无评论

Towards Precise 3D Human Pose Estimation with Multi-Perspective Spatial-Temporal Relational Transformers

Towards Precise 3D Human Pose Estimation with Multi-Perspect...

引用

International Joint Conference on Neural Networks (IJCNN)

作者： Jianbin Jiao Xina Cheng Weijie Chen Xiaoting Yin Hao Shi Kailun Yang School of Artificial Intelligence Xidian University China State Key Laboratory of Extreme Photonics and Instrumentation Zhejiang University China School of Robotics Hunan University China National Engineering Research Center of Robot Visual Perception and Control Technology Hunan University China

ISBN: (数字)9798350359312

ISBN: (纸本)9798350359329

3D human pose estimation captures the human joint points in three-dimensional space while keeping the depth information and physical structure. That is essential for applications that require precise pose information, such as humancomputer interaction, scene understanding, and rehabilitation training. Due to the challenges in data collection, mainstream datasets of 3D human pose estimation are primarily composed of multi-view video data collected in laboratory environments, which contains rich spatial-temporal correlation information besides the image frame content. Given the remarkable selfattention mechanism of transformers, capable of capturing the spatial-temporal correlation from multi-view video datasets, we propose a multi-stage framework for 3D sequence-to-sequence (seq2seq) human pose detection. Firstly, the spatial module represents the human pose feature by intra-image content, while the frame-image relation module extracts temporal relationships and 3D spatial positional relationship features between the multiperspective images. Secondly, the self-attention mechanism is adopted to eliminate the interference from non-human body parts and reduce computing resources. Our method is evaluated on Human3.6M, a popular 3D human pose detection dataset. Experimental results demonstrate that our approach achieves stateof-the-art performance on this dataset. The source code will be available at https://***/WUJINHUAN/3D-human-pose.

关键词： Training Three-dimensional displays Correlation Computational modeling Source coding Pose estimation Video sequences

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：