检索结果-内蒙古大学图书馆

arXiv 2021年

作者： Liu, Haozhe Wu, Haoqian Xie, Weicheng Liu, Feng Shen, Linlin 1Computer Vision Institute College of Computer Science and Software Engineering 2SZU Branch Shenzhen Institute of Artificial Intelligence and Robotics for Society 3National Engineering Laboratory for Big Data System Computing Technology 4Guangdong Key Laboratory of Intelligent Information Processing Shenzhen University Shenzhen 518060 China

The convolutional neural network (CNN) is vulnerable to degraded images with even very small variations (e.g. corrupted and adversarial samples). One of the possible reasons is that CNN pays more attention to the most discriminative regions, but ignores the auxiliary features when learning, leading to the lack of feature diversity for final judgment. In our method, we propose to dynamically suppress significant activation values of CNN by group-wise inhibition, but not fixedly or randomly handle them when training. The feature maps with different activation distribution are then processed separately to take the feature independence into account. CNN is finally guided to learn richer discriminative features hierarchically for robust classification according to the proposed regularization. Our method is comprehensively evaluated under multiple settings, including classification against corruptions, adversarial attacks and low data regime. Extensive experimental results show that the proposed method can achieve significant improvements in terms of both robustness and generalization performances, when compared with the state-of-the-art methods. Code is available at https://***/LinusWu/TENET_Training. Copyright © 2021, The Authors. All rights reserved.

关键词： Convolutional neural networks

来源：评论

学校读者我要写书评

暂无评论

Development and testing of an image transformer for explainable autonomous driving systems

arXiv

引用

arXiv 2021年

作者： Dong, Jiqian Chen, Sikai Zong, Shuya Chen, Tiantian Miralinaghi, Mohammad Labi, Samuel Lyles School of Civil Engineering Purdue University West LafayetteIN47907 United States Robotics Institute School of Computer Science Carnegie Mellon University PittsburghPA15213 United States Department of Industrial and System Engineering The Hong Kong Polytechnic University Hung Hom Kowloon Hong Kong

In the last decade, deep learning (DL) approaches have been used successfully in computer vision (CV) applications. However, DL-based CV models are generally considered to be black boxes due to their lack of interpretability. This black box behavior has exacerbated user distrust and therefore has prevented widespread deployment DLCV models in autonomous driving tasks even though some of these models exhibit superiority over human performance. For this reason, it is essential to develop explainable DL models for autonomous driving task. Explainable DL models can not only boost user trust in autonomy but also serve as a diagnostic approach to identify anydefects and weaknesses of the model during the system development phase. In this paper, we propose an explainable end-to-end autonomous driving system based on "Transformer," a state-of-the-art (SOTA) self-attention based model, to map visual features from images collected by onboard cameras to guide potential driving actions with corresponding explanations. The model achieves a soft attention over the global features of the image. The results demonstrate the efficacy of our proposed model as it exhibits superior performance (in terms of correct prediction of actions and explanations) compared to the benchmark model by a significant margin with lower computational cost. Copyright © 2021, The Authors. All rights reserved.

关键词： Computer vision

来源：评论

学校读者我要写书评

暂无评论

Object-oriented Map Exploration and Construction Based on Auxiliary Task Aided DRL

Object-oriented Map Exploration and Construction Based on Au...

引用

International Conference on Pattern Recognition

作者： Junzhe Xu Jianhua Zhang Shengyong Chen Honghai Liu College of Computer Science and Technology Zhejiang University of Technology Hangzhou China College of Computer Science and Technology Tianjin University of Technology Tianjin China State Key Laboratory of Robotics and System School of Mechanical Engineering and Automation Harbin Institute of Technology Shenzhen China

Environment exploration by autonomous robots through deep reinforcement learning (DRL) based methods has attracted more and more attention. However, existing methods usually focus on robot navigation to single or multiple fixed goals, while ignoring the perception and construction of external environments. In this paper, we propose a novel environment exploration task based on DRL, which requires a robot fast and completely perceives all objects of interest, and reconstructs their poses in a global environment map, as much as the robot can do. To this end, we design an auxiliary task aided DRL model, which is integrated with the auxiliary object detection and 6-DoF pose estimation components. The outcome of auxiliary tasks can improve the learning speed and robustness of DRL, as well as the accuracy of object pose estimation. Comprehensive experimental results on the indoor simulation platform AI2-THOR have shown the effectiveness and robustness of our method.

关键词： Navigation Object oriented modeling Pose estimation Semantics Object detection Reinforcement learning Robustness

来源：评论

学校读者我要写书评

暂无评论

Configuration synthesis of planar folded and common overconstrained spatial rectangular pyramid deployable truss units

引用

Chinese Journal of Aeronautics 2019年第7期32卷 1772-1787页

作者： Hui YANG Hongwei GUO Yan WANG Rongqiang LIU Zongquan DENG College of Electrical Engineering and Automation Anhui UniversityHefei230601China State Key Laboratory Robotics and System Harbin Institute of TechnologyHarbin150001China China Electronics Technology Group Corporation No.38Research Hefei230088China

Compared with non-overconstrained deployable units, overconstrained deployable units are widely used in space missions for their higher stiffness characteristics. Besides the performance of a three-step topological structural analysis and design of the rectangular pyramid deployable truss unit(PDTU), conducting a structural synthesis of an overconstrained deployable unit requires the determination of the relative position and direction of each kinematic axis. The structural synthesis of an overconstrained deployable unit is investigated based on screw theory and its topological structure. The possible overconstrained cases of the rectangular PDTUs are analyzed, and corresponding screw expressions are obtained. Thus, the rectangular PDTUs, which can be folded into a plane, are synthesized systemically, and a series of overconstrained rectangular PDTUs is obtained. Furthermore, the feasibility of the folded and deployed motions under one degree of freedom for those deployable units is verified in dynamical simulation by using ADAMS 2010.

关键词： Assembly SAR antenna Common overconstraint Deployable truss structure Screw theory Structural synthesis

来源：评论

学校读者我要写书评

暂无评论

Time-Optimal Control of Underactuated Wheeled Inverted Pendulum Vehicles Along Specified Paths

Time-Optimal Control of Underactuated Wheeled Inverted Pendu...

引用

International Conference on Advanced robotics and Mechatronics (ICARM)

作者： Yigao Ning Ming Yue Zenan Lin School of Automotive Engineering Dalian University of Technology Dalian China State Key Laboratory of Robotics and System Harbin Institute of Technology Harbin China

ISBN: (数字)9781728164793

ISBN: (纸本)9781728164809

This study proposes a time-optimal control method for a wheeled inverted pendulum (WIP) vehicle under specified paths. Firstly, the given path is described with a parametric form, and the kinematic model, velocity constraint and acceleration constraint are all parameterized with a same parameter as the geometric path. Meanwhile, the stabilization constraint of vehicle body is also formulated through the kinematic coupling relationship to take the underactuated characteristic into full consideration. Next, the time-optimal trajectory is constructed by a sequence of minimum acceleration and maximum acceleration trajectories based on phase-plane analysis, and the optimality is illustrated through rigorous analysis. Finally, a time-optimal trajectory tracking controller is designed for the WIP vehicle to realize the minimum-time motion, and numerical simulation results validate the effectiveness of the proposed method.

关键词： Acceleration Trajectory Mechatronics Kinematics Mobile robots Trajectory tracking

来源：评论

学校读者我要写书评

暂无评论

Review of surgical robotic systems for keyhole and endoscopic procedures:state of the art and perspectives

引用

Frontiers of Medicine 2020年第4期14卷 382-403页

作者： Yuyang Chen Shu’an Zhang Zhonghao Wu Bo Yang Qingquan Luo Kai Xu State Key Laboratory of Mechanical System and Vibration School of Mechanical EngineeringShanghai Jiao Tong UniversityShanghai 200240China RII Lab(Lab of Robotics Innovation and Intervention) UM-SJTU Joint InstituteShanghai Jiao Tong UniversityShanghai 200240China Department of Urology Shanghai Changhai Hospitalthe Second Military Medical UniversityShanghai 200433China Shanghai Lung Cancer Center Shanghai Chest HospitalShanghai Jiao Tong UniversityShanghai 200030China

Minimally invasive surgery,including laparoscopic and thoracoscopic procedures,benefits patients in terms of improved postoperative outcomes and short recovery *** challenges in hand–eye coordination and manipulation dexterity during the aforementioned procedures have inspired an enormous wave of developments on surgical robotic systems to assist keyhole and endoscopic procedures in the past *** paper presents a systematic review of the state-of-the-art systems,picturing a detailed landscape of the system configurations,actuation schemes,and control approaches of the existing surgical robotic systems for keyhole and endoscopic *** development challenges and future perspectives are discussed in depth to point out the need for new enabling technologies and inspire future researches.

关键词： surgical robots keyhole surgery endoscopic surgery

来源：评论

学校读者我要写书评

暂无评论

Autonomous WiFi Fingerprinting for Indoor Localization

Autonomous WiFi Fingerprinting for Indoor Localization

引用

IEEE/ACM International Conference on Cyber-Physical systems (ICCPS)

作者： Shilong Dai Liang He Xuebo Zhang Institute of Robotics and Automatic Information System (IRAIS) Nankai University Tianjin China Computer Science and Engineering University of Colorado Denver CO USA

ISBN: (数字)9781728155012

ISBN: (纸本)9781728155029

WiFi-based indoor localization has received extensive attentions from both academia and industry. However, the overhead of constructing and maintaining the WiFi fingerprint map remains a bottleneck for the wide-deployment of WiFi- based indoor localization systems. Recently, robots are adopted as the professional surveyor to fingerprint the environment autonomously. But the time and energy cost still limit the coverage of the robot surveyor, thus reduce its *** fill this need, we design an Autonomous WiFi Fingerprinting system, called AuF, which autonomously constructs the fingerprint database with improved time and energy efficiency. AuF first conduct an automatic initialization process in the target indoor environment, then constructs the WiFi fingerprint database in two steps: (i) surveying the site without sojourn, (ii) recovering unreliable signals in the database with two methods. We have implemented and evaluated AuF using a Pioneer 3-DX robot, on two sites of our 70 × 90m 2 Department building with different structures and deployments of access points (APs). The results show AuF finishes the fingerprint database construction in 43/51 minutes, and consumes 60/82 Wh on the two floors respectively, which is a 64%/71% and 61%/64% reduction when compared to traditional site survey methods, without degrading the localization accuracy.

关键词： Wireless fidelity Databases Buildings Robot kinematics Correlation Planning

来源：评论

学校读者我要写书评

暂无评论

NccFlow: Unsupervised learning of optical flow with non-occlusion from geometry

arXiv

引用

arXiv 2021年

作者： Wang, Guangming Ren, Shuaiqi Wang, Hesheng The Key Laboratory of System Control and Information Processing of Ministry of Education Department of Automation Institute of Medical Robotics Shanghai Jiao Tong University Shanghai200240 China The Key Laboratory of Marine Intelligent Equipment and System of Ministry of Education Shanghai Engineering Research Center of Intelligent Control and Management Shanghai Jiao Tong University Shanghai200240 China

Optical flow estimation is a fundamental problem of computer vision and has many applications in the fields of robot learning and autonomous driving. This paper reveals novel geometric laws of optical flow based on the insight and detailed definition of non-occlusion. Then, two novel loss functions are proposed for the unsupervised learning of optical flow based on the geometric laws of non-occlusion. Specifically, after the occlusion part of the images are masked, the flowing process of pixels is carefully considered and geometric constraints are conducted based on the geometric laws of optical flow. First, neighboring pixels in the first frame will not intersect during the pixel displacement to the second frame. Secondly, when the cluster containing adjacent four pixels in the first frame moves to the second frame, no other pixels will flow into the quadrilateral formed by them. According to the two geometrical constraints, the optical flow non-intersection loss and the optical flow non-blocking loss in the non-occlusion regions are proposed. Two loss functions punish the irregular and inexact optical flows in the non-occlusion regions. The experiments on datasets demonstrated that the proposed unsupervised losses of optical flow based on the geometric laws in non-occlusion regions make the estimated optical flow more refined in detail, and improve the performance of unsupervised learning of optical flow. In addition, the experiments training on synthetic data and evaluating on real data show that the generalization ability of optical flow network is improved by our proposed unsupervised approach. Copyright © 2021, The Authors. All rights reserved.

关键词： Computer vision

来源：评论

学校读者我要写书评

暂无评论

Closed-Form Modeling and Analysis of an XY Flexure-Based Nano-Manipulator

引用

Chinese Journal of Mechanical engineering 2018年第1期31卷 176-186页

作者： Yan-Ding Qin Xin Zhao Bijan Shirinzadeh Yan-Ling Tian Da-Wei Zhang Institute of Robotics and Automatic Information System (Tianjin Key Laboratory of Intelligent Robotics) Nankai University Robotics and Mechatronics Research Laboratory Department of Mechanical and Aerospace Engineering Monash University School of Mechanical Engineering Tianjin University

Flexure-based mechanisms are widely utilized in nano manipulations. The closed-form statics and dynamics modeling is difficult due to the complex topologies, the inevitable compliance of levers, the Hertzian contact interface, etc. This paper presents the closed-form modeling of an XY nano-manipulator consisting of statically indeterminate symmetric(SIS) structures using leaf and circular flexure hinges. Theoretical analysis reveals that the lever’s compliance, the contact stiffness, and the load mass have significant influence on the static and dynamic performances of the *** are conducted to verify the effectiveness of the established models. If no piezoelectric actuator(PEA) is installed, the influence of the contact stiffness can be eliminated. Experimental results show that the estimation error on the output stiffness and first natural frequency can reach 2% and 1.7%, respectively. If PEAs are installed, the contact stiffness shows up in the models. As no effective method is currently available to measure or estimate the contact stiffness, it is impossible to precisely estimate the performance of the overall system. In this case, the established closed-form models can be utilized to calculate the bounds of the performance. The established closed-form models are widely applicable in the design and optimization of planar flexure-based mechanisms.

关键词： Flexure-based mechanism Statically indeterminate structure Dynamics Lever mechanism Piezoelectric actuator

来源：评论

学校读者我要写书评

暂无评论

Study on the equilibrium of the assembling two-unit tensegrity structure

Study on the equilibrium of the assembling two-unit tensegri...

引用

作者： Luo, Ani Ji, Wanwei Liu, Heping Guo, Hongwei Liu, Rongqiang College of Mechanical and Electrical Engineering Harbin Engineering University Harbin150001 China State Key Laboratory of Robotics and System Harbin Institute of Technology Harbin150080 China

When the tensegrity units are assembled into the complex tensegrity structures, relations between structural and force parameters determine equilibrium of the tensegrity structures and contribute to more topological configurations, so the relations will be investigated here. Firstly, equilibrium equations of the tensegrity structure which two tensegrity units are assembled into are set up. Then analytical solutions of the equations are obtained by solving the equilibrium equations. Finally, with the analytical solutions, influences of the deflection angle inside the unit and number of the bars on other parameters are investigated. © 2019, Springer Nature Switzerland AG.

关键词： Analytical solutions Equilibrium equations Tensegrity structures Tensegrity unit

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：