检索结果-内蒙古大学图书馆

arXiv 2020年

作者： Chen, Xingyu Lu, Yue Wu, Zhengxing Yu, Junzhi Wen, Li State Key Laboratory of Management and Control for Complex Systems Institute of Automation Chinese Academy of Sciences Beijing100190 China School of Artificial Intelligence University of Chinese Academy of Sciences Beijing100049 China State Key Laboratory for Turbulence and Complex Systems Department of Mechanics and Engineering Science BIC-ESAT College of Engineering Peking University Beijing100871 China School of Mechanical Engineering and Automation Beihang University Beijing100191 China

Underwater robotic perception usually requires visual restoration and object detection, both of which have been studied for many years. Meanwhile, data domain has a huge impact on modern data-driven leaning process. However, exactly indicating domain effect, the relation between restoration and detection remains unclear. In this paper, we generally investigate the relation of quality-diverse data domain to detection performance. In the meantime, we unveil how visual restoration contributes to object detection in real-world underwater scenes. According to our analysis, five key discoveries are reported: 1) Domain quality has an ignorable effect on within-domain convolutional representation and detection accuracy;2) low-quality domain leads to higher generalization ability in cross-domain detection;3) low-quality domain can hardly be well learned in a domain-mixed learning process;4) degrading recall efficiency, restoration cannot improve within-domain detection accuracy;5) visual restoration is beneficial to detection in the wild by reducing the domain shift between training data and real-world scenes. Finally, as an illustrative example, we successfully perform underwater object detection with an aquatic robot. Copyright © 2020, The Authors. All rights reserved.

关键词： Object recognition

来源：评论

学校读者我要写书评

暂无评论

Optimal and stable control for two-player zero-sum game using adaptive dynamic programming 8

Optimal and stable control for two-player zero-sum game usin...

引用

8th IEEE Data Driven control and Learning systems Conference, DDCLS 2019

作者： Zhu, Liao Song, Ruizhuo Wei, Qinglai Yang, Zhanyu Zhao, Bo State Key Laboratory of Management and Control for Complex Systems Institute of Automation Chinese Academy of Sciences Beijing100190 China School of Automation and Electrical Engineering University of Science and Technology Beijing Beijing100083 China School of Systems Science Beijing Normal University Beijing100875 China

ISBN: (纸本)9781728114545

In this paper, an optimal and stable iteration learning scheme is developed for two-player zero-sum game (ZSG) in the discrete-time nonlinear systems. In the implementation of developed algorithm, the optimal and stable solution of Hamilton-Jacobi-Isaacs can be obtained based on adaptive dynamic programming (ADP). First, in order to obtain the optimal control policies, value iteration approach is employed with the proof of convergence given. Second, based on Lyapunov theory, an easy condition is proposed to get the stable control laws. Third, neural networks are used to implement the developed algorithm. Finally, a simulation example is included to verify the present method. © 2019 IEEE.

关键词： Neural networks

来源：评论

学校读者我要写书评

暂无评论

Research on Trajectory Planning and Motion control of 7-DoF Sawyer Manipulator

Research on Trajectory Planning and Motion Control of 7-DoF ...

引用

Digital Twins and Parallel Intelligence (DTPI), IEEE International Conference on

作者： Kun Luo Gang Xiong Yuxiao Wang Bin Hu Sheng Liu Shichao Chen The State Key Laboratory for Management and Control of Complex Systems Institute of Automation Chinese Academy of Sciences Beijng China School of Artificial Intelligence University of Chinese Academy of Sciences Beijng China Guangdong Engineering Research Center of 3D Printing and Intelligent Manufacturing Cloud Computing Center Chinese Academy of Sciences Beijng China Qingdao Academy of Intelligent Industries Chinese Academy of Sciences Beijng China

Regarding human-machine collaborative task, generated trajectories of a robot manipulator should be able to avoid joint limits and singularities, but the 6 Degrees of Freedom(DoF) or lower DoF manipulators encounter difficulties in many schemes. In this paper, research is implemented on the methods of trajectory planning and motion control on the newly human-machine collaborative 7-DoF Sawyer manipulator. We try to pay more attention to these approaches and find a better combination between trajectory planners and motion controllers for Sawyer manipulator. First of all, comparison is proposed between the methods about trajectory planning in task space and joint space, drawing a conclusion that the latter has the advantage of quickness and smoothness on trajectories in real time. Next, we choose to plan trajectory in the joint space by interpolating positions and orientations. Three approaches, i.e., trapezoidal velocity trajectory, cubic polynomial trajectory and basic spline trajectory, are applied to interpolate positions. Experimental results show that it results in the smoothest trajectory using basic spline trajectory planning. We also compare linear and nonlinear orientation interpolations to address the angle wrapping issue, which leads to difficulties in smooth orientation interpolation between several waypoints. Combining polynomial trajectory planning methods with the proportional-derivative (PD) motion controller and the computed-torque motion controller, we focus on the performance of the two different combinations. It is found that trajectories with computed-torque controller have shorter path with a congruous convergence on joint position, and more accurate trajectories with less overshoots in position than others.

关键词： Interpolation Trajectory planning Collaboration Aerospace electronics Manipulators Trajectory Task analysis

来源：评论

学校读者我要写书评

暂无评论

3D Printer Optical Detection System Based On DLP Projection Technology

3D Printer Optical Detection System Based On DLP Projection ...

引用

Chinese Automation Congress (CAC)

作者： Guoliang Wu Zhen Shen Xiuqin Shang Huaiyu Wu Gang Xiong Jing Yang The State Key Laboratory for Management and Control of Complex Systems Institute of Automation Chinese Academy of Sciences Beijing China

In order to solve the printing failure caused by the projection part of the light source in the 3D printing process, this paper proposes a system adjustment scheme. By analyzing the polymerization state of the photosensitive resin, detecting the intensity of ultraviolet light, extracting the outline of the picture, and comparing the similarity of the picture, the optical detection system can find out the cause of the printing failure and provide us with advice on how to adjust the system to ensure that the system can continue to be normal. In the following, the authors proposed the detection process of the system, and simulated the experiment to confirm whether the ultraviolet light irradiated on the photosensitive resin is the same as the image in the slice file.

关键词： Resins Lenses Light sources Three-dimensional printing Printers Optical sensors

来源：评论

学校读者我要写书评

暂无评论

Reinforcement Learning based Lane Change Decision-Making with Imaginary Sampling

Reinforcement Learning based Lane Change Decision-Making wit...

引用

IEEE Symposium Series on Computational Intelligence (SSCI)

作者： Dong Li Dongbin Zhao Qichao Zhang The State Key Laboratory of Management and Control for Complex Systems Institute of Automation Chinese Academy of Sciences Beijing China

ISBN: (数字)9781728124858

ISBN: (纸本)9781728124865

This paper investigates the autonomous driving lane change problem with reinforcement learning methods. A two-stage control method is proposed which includes a decision-making module computing the high-level lane change action and a lateral control module outputting the low-level steering angle. Due to the decoupling system framework, the proposed controller can flexibly cancel the previous improper lane change command. To improve the data efficiency in the lateral control module, the Gaussian process is employed to modeling the local time-interval system models. These models are used to generate imaginary samples and facilitate the reinforcement learning training process. The lateral control experiments show that the imaginary samples can improve the data efficiency and speed up the training process. Finally, the two-lane scene and three-lane scene experiments validate the effectiveness of the proposed two-stage lane change method.

关键词： Decision making Learning (artificial intelligence) Gaussian processes Uncertainty Estimation Autonomous vehicles Roads

来源：评论

学校读者我要写书评

暂无评论

Spread-gram: A spreading-Activation schema of network structural learning

arXiv

引用

arXiv 2019年

作者： Bai, Jie Li, Linjing Zeng, Daniel State Key Laboratory of Management and Control for Complex Systems Institute of Automation Chinese Academy of Sciences Beijing China

Network representation learning has exploded recently. However, existing studies usually reconstruct networks as sequences or matrices, which may cause information bias or sparsity problem during model training. Inspired by a cognitive model of human memory, we propose a network representation learning scheme. In this scheme, we learn node embeddings by adjusting the proximity of nodes traversing the spreading structure of the network. Our proposed method shows a significant improvement in multiple analysis tasks based on various real-world networks, ranging from semantic networks to protein interaction networks, international trade networks, human behavior networks, etc. In particular, our model can effectively discover the hierarchical structures in networks. The well-organized model training speeds up the convergence to only a small number of iterations, and the training time is linear with respect to the edge numbers. Copyright © 2019, The Authors. All rights reserved.

关键词： Embeddings

来源：评论

学校读者我要写书评

暂无评论

Design and Validation of an Asymmetric Bowden-Cable-Driven Series Elastic Actuator

Design and Validation of an Asymmetric Bowden-Cable-Driven S...

引用

IEEE International Conference on Robotics and Biomimetics

作者： Ning Sun Long Cheng Lin Tian Zeng-Guang Hou Min Tan State Key Laboratory of Management and Control for Complex Systems Institute of Automation Chinese Academy of Sciences Beijing China

Weight, size and torque control for hand exoskeleton robots are challenging due to lack of small and compact bidirectional torque actuators. In this paper, an asymmetric Bowden-cable-driven series elastic actuator (ABLE-SEA) for hand exoskeleton robots was proposed based on the fact that the required torques for finger movements are asymmetric. Compared to series elastic actuator and transmission system that were designed separately, the elastic elements of ABLESEA were placed in transmission parts to connect series elastic actuator with transmission system, which makes system compact. ABLE-SEA is 71 mm × 19.5mm × 20mm in dimension and weighs 30g except motor which is placed remotely. The dynamic model of ABLE-SEA was established, and feedback proportional-derivative (PD) control plus a feed-forward term was used to track the reference torque for the ABLE-SEA. The reference torque tracking test at different frequencies was performed at the developed prototype. Meanwhile, the peak torque of ABLE-SEA was tested. The experimental results verified that the torque bandwidth of the proposed series elastic actuator could reach 4Hz, and the peak torque of ABLE-SEA could reach 0.3Nm, which meet the design requirement.

关键词：

来源：评论

学校读者我要写书评

暂无评论

A Building Evacuation Model Integrated with Multiple Algorithms

A Building Evacuation Model Integrated with Multiple Algorit...

引用

Chinese Automation Congress (CAC)

作者： Xiaoyu Zhao Xiwei Liu Xiaoyan Gong Sifeng Jing Jian He The State Key Laboratory of Management and Control for Complex Systems Institute of Automation Beijing University of Technology BeiJing China

Due to the increasing frequency of terrorist attacks all over the world, we designed an emergency evacuation model under the situation of the terrorist attacks. This model uses the Ant Colony Optimization (ACO) as a macro solution, the Cellular Automata (CA) as a micro-control method, and the Queuing Theory Model (QTM) to solve the queuing problem of escape exits. Based on the model we built, the optimized evacuation plan under different conditions were obtained.

关键词： Automata Queueing analysis Terrorism Ant colony optimization Analytical models Standards Buildings

来源：评论

学校读者我要写书评

暂无评论

Simplified Space Based Neural Architecture Search

Simplified Space Based Neural Architecture Search

引用

IEEE Symposium Series on Computational Intelligence (SSCI)

作者： Zixiang Ding Yaran Chen Nannan Li Dongbin Zhao State Key Laboratory of Management and Control for Complex Systems Institute of Automation Chinese Academy of Sciences Beijing China

ISBN: (数字)9781728124858

ISBN: (纸本)9781728124865

In this paper, we propose Simplified Space based Neural Architecture Search (SSNAS), an efficient approach for automatic architecture search. Inspired by the popular convolutional neural networks which have strong capability for feature extraction with small convolutional kernels, we design a simplified search space of convolutional operations with small kernel and construct a large model based on it. Furthermore, we use an Long Short-Term Memory (LSTM) to sample child model from the large model in a way of selective activation. The probability distribution of selective activation is obtained by training the LSTM with reinforcement learning for maximizing the expected reward of selected child model on the validation set. Moreover, the trained weights are saved in a large model which concludes all child models. Each weight of new sampled child model will be restored when being selected by another one again instead of training from scratch so that SSNAS can greatly reduce the expensive computation than traditional approaches for automatic model design. Extensive experiments on CIFAR-10, ImageNet and VOC show that the proposed approach excels in discovering transferable architectures with high performance.

关键词： Computer architecture Microprocessors Computational modeling Search problems Estimation Training Kernel

来源：评论

学校读者我要写书评

暂无评论

Object Pose Estimation Based on RGB-D Sensor for Cooperative Spray Painting Robot

Object Pose Estimation Based on RGB-D Sensor for Cooperative...

引用

IEEE International Conference on Cyber Technology in Automation, control, and Intelligent systems

作者： Zhe Wang Fengshui Jing Junfeng Fan Zhaoyang Liu Yunong Tian Zishu Gao State Key Laboratory of Management and Control for Complex Systems Institute of Automation Chinese Academy of Sciences Beijing China

ISBN: (数字)9781728107707

ISBN: (纸本)9781728107714

For human-robot cooperative spray painting robot, offline programming based on predefined model of the unpainted object is a robust and efficient method for trajectory generation. To apply the programmed trajectory on the unpainted object, the relative pose between the object and the predefined model needs to be acquired. Nevertheless, acquiring an accurate estimation of the pose in spray painting setting remains a problem. To address this, a RGB-D pose estimation system based on deep learning and iterative closest point (ICP) alignment is proposed in this paper. The perception module of this system is RGB-D sensor. The RGB-D image of the object is segmented using Fully Convolutional Network (FCN) with RGB-D input. The resulting segmented point cloud is aligned with the model candidates using ICP algorithm to estimate the pose of the object. It is validated in the experiments that the proposed system and methods are effective and robust.

关键词： Painting Robot sensing systems Solid modeling Three-dimensional displays Pose estimation Image segmentation Iterative closest point algorithm

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：