检索结果-内蒙古大学图书馆

DG-NBV: A Cognitive Framework for Direct Generation of Next Best View in Continuous View Space

IEEE Transactions on Cognitive and Developmental Systems 2024年

作者： Liu, Zhicheng Cao, Zhiqiang Li, Jianjie Guan, Peiyu Yu, Junzhi Chinese Academy of Sciences State Key Laboratory of Multimodal Artificial Intelligence Systems Institute of Automation Beijing100190 China University of Chinese Academy of Sciences School of Artificial Intelligence Beijing100049 China Peking University Department of Advanced Manufacturing and Robotics BIC-ESAT College of Engineering Beijing100871 China

The next best view (NBV) plays an important role in the three-dimensional object reconstruction. Existing NBV methods mainly adopt the generate-and-test strategy to select the best view in the discrete view space. This affects the adaptation to diverse objects. To solve this problem, a new NBV paradigm to directly generate the next best view in the continuous view space according to the input point cloud is proposed. Specifically, a point cloud feature extraction module with learnable view and position tokens is presented. These tokens are added upon the neighborhood and the position features of the point cloud to fully mine global contextual information, enhancing the feature representation. The predicted view from the proposed network is linked to a pre-trained view evaluation network. By duplicating this prediction view and then concatenating the duplication result with the extracted point cloud feature, the evaluation network is endowed with the ability to evaluate arbitrary views. Take the evaluation score of the evaluation network corresponding to the predicted view as the supervision signal, the network is trained. In this way, an effective solution of the next best view selection in the continuous view space is obtained. Accordingly, the adaptability to different objects is reinforced. Experimental results on the ShapeNet and MIT CSAIL datasets demonstrate the effectiveness of the proposed method. © 2024 IEEE.

关键词： Feature extraction

来源：评论

学校读者我要写书评

暂无评论

Learning to Turn: Deformation Control of a Novel Flexible Fishtail Actuated by Artificial Muscles

Learning to Turn: Deformation Control of a Novel Flexible Fi...

引用

IEEE International Conference on Cyborg and Bionic Systems (CBS)

作者： Junwen Gu Yukai Feng Zhengxing Wu Jian Wang Junzhi Yu Key Laboratory of Cognition and Decision Intelligence for Complex Systems Institute of Automation Chinese Academy of Sciences Beijing China School of Artificial Intelligence University of Chinese Academy of Sciences Beijing China Department of Advanced Manufacturing and Robotics State Key Laboratory for Turbulence and Complex Systems College of Engineering Peking University Beijing China

ISBN: (数字)9798350388039

ISBN: (纸本)9798350388046

Achieving agile, fish-like turning remains a critical challenge in the development of robotic fish. The flexible deformation of a fish’s body significantly contributes to its turning ability. Consequently, researchers have devoted substantial efforts to enhancing the compliance and agility of robotic fish. This paper presents a novel flexible fishtail, inspired by the musculoskeletal structure of real fish, which can control the deformation of its flexible components during turns, improving the turning performance of robotic fish. The fishtail primarily consists of a carbon fiber plate with attached artificial muscles. These muscles can contract and stretch, causing the carbon fiber tail to bend and deform. To enable precise deformation control, a dynamic model of the fishtail is established based on Hamilton’s principle. Subsequently, deep reinforcement learning methods such as deep deterministic policy gradient and soft actor-critic are employed to develop a deformation controller under joint deflection. Cross-validation ensures the controller’s accuracy and robustness. The experimental results demonstrate that this controller facilitates the deformation control of the flexible tail across various turning amplitudes and speeds, allowing the fishtail to exhibit adjustable compliance or resistance to fluids during turning motion. Consequently, this enhances the agility of the robotic fish.

关键词： Artificial muscles Deformation Tail Learning (artificial intelligence) Muscles Turning Fish Robustness Carbon Robots

来源：评论

学校读者我要写书评

暂无评论

Real-Time Stereo Matching Networks Based on Multi-Modal and Sparse-Dense Fusion

SSRN

引用

SSRN 2023年

作者： Zhang, Xi Wu, Xiaojun Liang, Guoyuan School of Mechanical Engineering and Automation Harbin Institute of Technology Shenzhen Shenzhen518055 China Guangdong Provincial Key Laboratory of Robotics and Intelligent System Shenzhen Institute of Advanced Technology Chinese Academy of Sciences Shenzhen518055 China

Binocular depth estimation models based on deep learning are difficult to balance accuracy and efficiency, and often occupy a lot of memory during inference. This paper proposes a new, concise and easily scalable model architecture based on disparity iterative optimization, Fusion-Stereo, which can achieve real-time inference speed and low memory occupation on the basis of ensuring excellent prediction accuracy. Based on the observation of different modal information of image texture feature, matching similarity feature and disparity feature, we propose a new disparity residual prediction module based on multi-modal information coding and fusion to help the adaptive propagation of reliable disparity information in spatial dimension and disparity dimension. In order to improve the reliability of depth prediction, we design an efficient fusion mechanism between Fusion-Stereo and external priori data sources, and introduce sparse disparity priori into the model to facilitate the dense disparity prediction, while ensuring that the inference speed and memory occupation are basically unchanged. © 2023, The Authors. All rights reserved.

关键词： Stereo vision

来源：评论

学校读者我要写书评

暂无评论

TRUNK-BRANCH CONTRASTIVE NETWORK WITH MULTI-VIEW DEFORMABLE AGGREGATION FOR MULTI-VIEW ACTION RECOGNITION

arXiv

引用

arXiv 2025年

作者： Yang, Yingyuan Liang, Guoyuan Wang, Can Wu, Xiaojun Shenzhen Institutes of Advanced Technology Chinese Academy of Sciences Shenzhen518055 China University of Chinese Academy of Sciences Beijing10049 China Guangdong Provincial Key Laboratory of Robotics and Intelligent System Shenzhen Institutes of Advanced Technology Chinese Academy of Sciences Shenzhen518055 China School of Mechanical Engineering and Automation Harbin Institute of Technology at Shenzhen Shenzhen518055 China

Multi-view action recognition aims to identify actions in a given multi-view scene. Traditional studies initially extracted refined features from each view, followed by implemented paired interaction and integration, but they potentially overlooked the critical local features in each view. When observing objects from multiple perspectives, individuals typically form a comprehensive impression and subsequently fill in specific details. Drawing inspiration from this cognitive process, we propose a novel trunk-branch contrastive network (TBCNet) for RGB-based multi-view action recognition. Distinctively, TBCNet first obtains fused features in the trunk block and then implicitly supplements vital details provided by the branch block via contrastive learning, generating a more informative and comprehensive action representation. Within this framework, we construct two core components: the multi-view deformable aggregation (MVDA) and the trunk-branch contrastive learning. MVDA employed in the trunk block effectively facilitates multi-view feature fusion and adaptive cross-view spatio-temporal correlation, where a global aggregation module (GAM) is utilized to emphasize significant spatial information and a composite relative position bias (CRPB) is designed to capture the intra- and cross-view relative positions. Moreover, a trunk-branch contrastive loss is constructed between aggregated features and refined details from each view. By incorporating two distinct weights for positive and negative samples, a weighted trunk-branch contrastive loss is proposed to extract valuable information and emphasize subtle inter-class differences. The effectiveness of TBCNet is verified by extensive experiments on four datasets including NTU-RGB+D 60, NTU-RGB+D 120, PKU-MMD, and N-UCLA dataset. Compared to other RGB-based methods, our approach achieves state-of-the-art performance in cross-subject and cross-setting protocols. Copyright © 2025, The Authors. All rights reserved.

关键词： Contrastive Learning

来源：评论

学校读者我要写书评

暂无评论

Modeling Neural Representation of Social Recognition Information with Recurrent Neural Networks 17

Modeling Neural Representation of Social Recognition Informa...

引用

17th International Conference on Complex Medical Engineering, CME 2023

作者： Song, Fei Wang, Xinnian Zhan, Yang Tang, Fengzhen Si, Bailu State Key Laboratory of Robotics Shenyang Institute of Automation 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 School of Life Sciences Division of Life Science and Medicine University of Science and Technology of China Hefei230026 China Shenzhen Institutes of Advanced Technology Chinese Academy of Sciences Shenzhen518055 China School of Systems Science Beijing Normal University Beijing100091 China Chinese Institute for Brain Research Beijing102206 China

ISBN: (纸本)9798350316117

Social recognition, the ability to differentiate between familiar and unfamiliar individuals, is crucial for animals' survival and reproduction. Neuroscience research has predominantly focused on investigating the brain regions and circuits that regulate social recognition. However, a lingering question pertains to the precise mechanisms underlying the social recognition process. To address this question, we employed a recurrent neural network (RNN) to determine whether an agent can recognize familiar individuals without involving higher-level decision-making processes. This approach takes into consideration that animal behavioral performance may not directly mirror complex cognitive functions. The RNN was trained using the full-FORCE learning paradigm. Our simulation results revealed three distinct subsets of neurons with specific functions: one dedicated to representing novel individuals and two others representing the same social individual in both unfamiliar and familiar states. We propose a hypothesis regarding the potential memory computing processes involved in these operations. © 2023 IEEE.

关键词： Animals

来源：评论

学校读者我要写书评

暂无评论

Sequential Manipulation Planning for Over-Actuated Unmanned Aerial Manipulators

Sequential Manipulation Planning for Over-Actuated Unmanned ...

引用

IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS)

作者： Yao Su Jiarui Li Ziyuan Jiao Meng Wang Chi Chu Hang Li Yixin Zhu Hangxin Liu National Key Laboratory of General Artificial Intelligence Beijing Institute for General Artificial Intelligence (BIGAI) Department of Advanced Manufacturing and Robotics College of Engineering Peking University Department of Automation Tsinghua University Institute for Artificial Intelligence Peking University

We investigate the sequential manipulation planning problem for unmanned aerial manipulators (UAMs). Unlike prior work that primarily focuses on one-step manipulation tasks, sequential manipulations require coordinated motions of a UAM's floating base, the manipulator, and the object being manipulated, entailing a unified kinematics and dynamics model for motion planning under designated constraints. By leveraging a virtual kinematic chain (VKC)-based motion planning framework that consolidates components' kinematics into one chain, the sequential manipulation task of a UAM can be planned as a whole, yielding more coordinated motions. Integrating the kinematics and dynamics models with a hierarchical control framework, we demonstrate, for the first time, an over-actuated UAM achieves a series of new sequential manipulation capabilities in both simulation and experiment.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Non-cooperative Stochastic Target Encirclement by Anti-synchronization Control via Range-only Measurement

Non-cooperative Stochastic Target Encirclement by Anti-synch...

引用

IEEE International Conference on robotics and automation (ICRA)

作者： Fen Liu Shenghai Yuan Wei Meng Rong Su Lihua Xie Guangdong Provincial Key Laboratory of Intelligent Decision and Cooperative Control School of Automation Guangdong University of Technology Guangzhou the Centre for Advanced Robotics Technology Innovation (CARTIN) School of Electrical and Electronic Engineering Nanyang Technological University Singapore Singapore

This paper investigates the stochastic moving target encirclement problem in a realistic setting. In contrast to typical assumptions in related works, the target in our work is non-cooperative and capable of escaping the circle containment by boosting its speed to maximum for a short duration. In extreme conditions, where GPS signals are not available, weight restrictions are present, and ground guidance is absent, the agents can rely solely on their onboard single-modality perception tools to measure the distances to the target. The distance measurement allows for creating a position estimator by providing a target position-dependent variable. Furthermore, the construction of the unique distributed anti-synchronization controller (DASC) can guarantee that the two agents track and encircle the target swiftly. The convergence of the estimator and controller is rigorously evaluated using the Lyapunov technique. A real-world UAV-based experiment is conducted to illustrate the performance of the proposed methodology in addition to a simulated Matlab numerical sample. Our video demonstration can be found in the URL https://***/EDVLvP-bk8M.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Synthesis of Decoherence-Free Modes in Linear Quantum Passive Systems via Robust Pole Placement

Synthesis of Decoherence-Free Modes in Linear Quantum Passiv...

引用

IEEE International Conference on Systems, Man and Cybernetics

作者： Zibo Miao Yu Pan Qing Gao Guangdong Key Laboratory of Intelligent Morphing Mechanisms School of Mechanical Engineering and Automation Adaptive Robotics Harbin Institute of Technology Shenzhen China State Key Laboratory of Industrial Control Technology The Institute of Cyber-Systems and Control College of Control Science and Engineering Zhejiang University Hangzhou China The School of Automation Science and Electrical Engineering Beijing Advanced Innovation Center for Big Data and Brain Computing Beihang University Beijing China

ISBN: (数字)9781665410205

ISBN: (纸本)9781665410212

In this paper we extend our previous research on coherent observer-based pole placement approach to study the synthesis of robust decoherence-free (DF) modes for linear quantum passive systems, which is aimed at preservation of quantum information. In particular, DF modes can be generated by placing the poles on the imaginary axis via a coherent feedback design scheme, and these modes can further be simultaneously made robust against perturbations to the system parameters by minimizing the condition number associated with imaginary poles. We develop explicit algebraic conditions for the existence of such a coherent quantum controller, with the corresponding deign procedure provided. Examples are given to illustrate the process of tuning the DF modes towards perfect robustness via the proposed pole placement technique.

关键词： Perturbation methods Design methodology Quantum optics Robustness Feedback control Closed loop systems Tuning Cybernetics

来源：评论

学校读者我要写书评

暂无评论

Improve the Performance of Active Knee Prosthesis During Stairs Ascent by Using the Leading Reference Control System

Improve the Performance of Active Knee Prosthesis During Sta...

引用

IEEE International Conference on Cyborg and Bionic Systems (CBS)

作者： Qinpeng Huang Wenhao Wei Bojun Huang Hui Wang Pingao Huang Guanglin Li School of Electronic Engineering and Automation Guilin Key Laboratory of Intelligence Integrated Automation in Guangxi Universities Guilin University of Electronic Technology Guangxi China CAS Key Laboratory of HumanMachine Intelligence-Synergy Systems Shenzhen Institutes of Advanced Technology (SIAT) Chinese Academy of Sciences (CAS) and the SIAT Branch Shenzhen Institute of Artificial Intelligence and Robotics for Society Shenzhen China

Strong coupling between the human and machine exists in many cyborg and bionic systems (CBS). Some input signals for control system barely change after the outputs work. In this situation, it is hard for the controller to calculate the required reference output of system. For example, in the stance phase of stairs ascent gait using an active knee prosthesis (AKP), most AKPs use the Inertial Measurement Unit (IMU) signal and the Ground Reaction Forces (GRF) signal to control the output of knee joint torque. However, these two signals hardly change unless the controlled object (joint torque) is already close to the actual requirement. Therefore, this study introduces a leading reference control system (LRCS) that uses an overshoot reference value to make the AKP more effective in providing positive energy in the abovementioned situation. In addition, the LRCS with different parameters was evaluated through the experiments of stairs ascent with AKP. Results show that LRCS with the suitable parameter can provide greater torque output when going upstairs, making the subject go upstairs easier.

关键词： Knee Couplings Torque Measurement units Inertial navigation Stairs Control systems

来源：评论

学校读者我要写书评

暂无评论

Data-driven modeling the nonlinear backlash of steerable endoscope under a large deflection cannulation in ERCP surgery

Data-driven modeling the nonlinear backlash of steerable end...

引用

2021 IEEE International Conference on Real-Time Computing and robotics, RCAR 2021

作者： Jiang, Wei Yu, Tao He, Xiao Yang, Yongming Wang, Zhidong Liu, Hao Shenyang Institute of Automation Institutes for Robotics and Intelligent Manufacturing Chinese Academy of Sciences State Key Laboratory of Robotics Shenyang110016 China University of Chinese Academy of Sciences Beijing100049 China Key Laboratory of Minimally Invasive Surgical Robot Liaoning Province Shenyang110016 China Key Laboratory of Minimally Invasive Surgical Liaoning Province Shenyang110016 China Department of Advanced Robotics Chiba Institute of Technology Narashino Chiba275-0016 Japan

ISBN: (纸本)9781665436786

Unexpected nonlinear backlash performances of two degrees of freedom (DoFs) tendon sheath mechanism (TSM) have serious impacts on the precise cannulation in ERCP (Endoscopic Retrograde Cholangio-Pancreatography) surgery. We proposed to model backlash performance of two DoFs steerable endoscope by using LSTM (Long Short-Term Memory) network. Trajectory following and orientation tasks are performed after the backlash model has been considered into the control strategy. Experimental results show that the LSTM backlash model can accurately describe the non-linearities of backlash, and significantly reduce the positioning and orientation deviation of the steerable endoscope under large deflection. The proposed backlash modeling method can be extended to be applied to bronchial, urethral and other natural orifice interventional tasks that work under a large deflection. © 2021 IEEE.

关键词： Endoscopy

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：