检索结果-内蒙古大学图书馆

Enhanced spatio-temporal motion prediction using transformer-augmented graph convolutional networks

INTERNATIONAL JOURNAL OF machine LEARNING AND CYBERNETICS 2025年 1-17页

作者： Ren, Ziliang Jin, Miaomiao Qin, Yong Gao, Xiangyang Zhang, Qieshi Dongguan Univ Technol Sch Comp Sci & Technol Dongguan 523808 Guangdong Peoples R China Chinese Acad Sci Shenzhen Inst Adv Technol Key Lab Human Machine Intelligence Synergy Syst Shenzhen 518055 Guangdong Peoples R China

Mining temporal and spatial correlations from observed data is a crucial challenge in human motion prediction. Traditional methods mainly rely on artificially designed features, which cannot accurately capture complex motion patterns and often ignore the reliability of motion. In recent years, deep learning methods have made significant progress in spatiotemporal representation, especially when dealing with human motion prediction tasks. Although temporal and spatial features have different characteristics, existing methods often fail to fully consider the temporal and spatial features. Correlation leads to inaccurate modeling of spatiotemporal information. In this paper, we propose a novel branch network based on the Transformer and augmented graph convolutional network (TAGCN). This method innovatively combines the global dependency modeling capability of the Transformer model with the local feature extraction capability of the graph convolution network. The Transformer model effectively captures long-distance dependencies in time series through a self-attention mechanism and is able to identify complex temporal patterns. At the same time, GCN models the spatial dependence between joints through a graph structure and can effectively extract local spatial features. Our method significantly improves the expressive ability of motion information by learning temporal and spatial features simultaneously, making it richer and more representative. In addition, we design an improved loss function to further improve the accuracy of spatiotemporal prediction. Experimental results show that our method achieves state-of-the-art prediction performance on human3.6M, CMU Mocap and 3DPW datasets, significantly outperforming existing mainstream methods.

关键词： human motion prediction Spatio-temporal modelling Transformer Graph convolutional network

来源：评论

学校读者我要写书评

暂无评论

Autonomous 3D Self-Sensing Hybrid Membrane Actuator for Interactive Communicating

引用

ACS APPLIED MATERIALS & INTERFACES 2025年第22期17卷 2025 Apr 29页

作者： Hou, Yuanyuan Huang, Jiaxin Ma, Hao Li, Qingsong Xiang, Zerong Qian, Jiasheng Li, Guanglin Tai, Yanlong Xia, Ru Zhu, Shanshan Anhui Univ Sch Chem & Chem Engn Key Lab Environm Friendly Polymer Mat Anhui Prov Hefei 230601 Anhui Peoples R China Chinese Acad Sci Shenzhen Inst Adv Technol CAS Key Lab Human Machine Intelligence Synergy Sys Shenzhen 518055 Peoples R China

The advancement of intelligent soft actuators is progressively emphasizing the incorporation of environmental sensing capability to actuation, thereby enhancing the adaptability and interactivity of artificial systems. In the current situation where the sensing and actuation functions of soft actuators are generally separated, this work proposes an autonomous three-dimensional (3D) noncontact sensory actuator (NSA), based on the coupling of '' dielectric polarization-electrothermal conversion-thermal actuation '' triple effects. Specifically, the NSA hybrid membrane is composed of multiple interpenetrating networks, including a boron nitride nanosheet (BNNS) dielectric network for electrostatic field sensing and polarization, a silver nanowires (AgNWs) percolation network for dielectric enhancement and electrothermal conversion, and thermally contracted shape memory fiber (SMF) and thermally expanded polydimethylsiloxane (PDMS) networks for directional actuation. Based on the principle of electrostatic field and dielectric polarization, the SMF/BNNS composite (SMF-BN) fibrous membrane can logically sense the noncontact 3D motion, static/dynamic state of external objects, and distinguish material categories. Subsequently, the output sensing potential facilitates the built-in AgNWs nanonetwork heater to trigger electrothermal actuation of NSA. Lastly, as a biomimetic tongue, the autonomous noncontact "sensing-decision-actuating" of NSA is verified by seamless energy conversion in the process of sensing "prey" approaching and capturing. The proposed sensory actuator would facilitate multimodal integration for future wearable and human-machine-environment interaction technologies.

关键词： 3D noncontact sensing soft sensory actuator energy conversion human-machine-environment interaction biomimetic device

来源：评论

学校读者我要写书评

暂无评论

A novel spatial feature extraction method based on high-density sEMG for complex hand movement recognition

引用

BIOMEDICAL SIGNAL PROCESSING AND CONTROL 2025年 103卷

作者： Peng, Boxing Zhang, Haoshi Li, Xiangxin Li, Guanglin Chinese Acad Sci Shenzhen Inst Adv Technol SIAT CAS Key Lab Human Machine Intelligence Synergy Sys Shenzhen 518055 Peoples R China Univ Chinese Acad Sci Beijing 100049 Peoples R China Shenzhen Inst Artificial Intelligence & Robot Soc SIAT Branch Shenzhen 518055 Peoples R China

Surface electromyography pattern recognition (sEMG-PR) based prostheses control system is beneficial for amputees to restore their upper limb function. However, identifying complex human movements from surface electromyography (sEMG) remains challenging for electromyography pattern recognition system because complex hand and wrist movements involve more subtle muscular regional activities that may not be captured by the classic independent channel extraction method with time-domain (TD) features. In this study, a novel spatial feature extraction method, Multichannel linear descriptors based Block field method (MLD-BFM), was proposed to improve the movement classification performance by extracting the spatial information of high-density sEMG (HD-sEMG) signals. In this method, a 64-channel HD-sEMG system was used, BFM was designed to segment regions from channel grids, and MLD was employed as spatial feature set extracted from each region. The effectiveness of the proposed method was evaluated to identify 28 classes of hand and wrist movements with comparison to six well-known feature extraction methods. The results showed that the proposed spatial feature extraction method achieved the classification error rates (CE) of 3.59%, which is the lowest for identifying 28 classes of movements, and had the lowest CEs across all four classifiers employed in this study. Furthermore, the proposed method showed greatly enhancement in combined movement recognition with a CE of 4.94%, which is the lowest across all feature extraction methods. The work of this study may provide a promising approach to extract the spatial information of HD-sEMG from different muscle regions to classify the complex hand movements.

关键词： High-density electromyography Spatial feature extraction Pattern recognition Myoelectric control

来源：评论

学校读者我要写书评

暂无评论

Masked cosine similarity prediction for self-supervised skeleton-based action representation learning

引用

PATTERN ANALYSIS AND APPLICATIONS 2025年第2期28卷 1-17页

作者： Ren, Ziliang Liu, Ronggui Qin, Yong Gao, Xiangyang Zhang, Qieshi Dongguan Univ Technol Sch Comp Sci & Technol Dongguan 523808 Guangdong Peoples R China Chinese Acad Sci Shenzhen Inst Adv Technol CAS Key Lab Human Machine Intelligence Synergy Sys Shenzhen 518055 Guangdong Peoples R China

Skeleton-based human action recognition faces challenges owing to the limited availability of annotated data, which constrains the performance of supervised methods in learning representations of skeleton sequences. To address this issue, researchers have introduced self-supervised learning as a method of reducing the reliance on annotated data. This approach exploits the intrinsic supervisory signals embedded within the data itself. In this study, we demonstrate that considering relative positional relationships between joints, rather than relying on joint coordinates as absolute positional information, yields more effective representations of skeleton sequences. Based on this, we introduce the Masked Cosine Similarity Prediction (MCSP) framework, which takes randomly masked skeleton sequences as input and predicts the corresponding cosine similarity between masked joints. Comprehensive experiments show that the proposed MCSP self-supervised pre-training method effectively learns representations in skeleton sequences, improving model performance while decreasing dependence on extensive labeled datasets. After pre-training with MCSP, a vanilla transformer architecture is employed for fine-tuning in action recognition. The results obtained from six subsets of the NTU-RGB+D 60, NTU-RGB+D 120 and PKU-MMD datasets show that our method achieves significant performance improvements on five subsets. Compared to training from scratch, performance improvements are 9.8%, 4.9%, 13%, 11.5%, and 3.6%, respectively, with top-1 accuracies of 92.9%, 97.3%, 89.8%, 91.2%, and 96.1% being achieved. Furthermore, our method achieves comparable results on the PKU-MMD Phase II dataset, achieving a top-1 accuracy of 51.5%. These results are competitive without the need for intricate designs, such as multi-stream model ensembles or extreme data augmentation. The source code of our MOSP is available at https://***/skyisyourlimit/MCSP.

关键词： Skeleton-based action recognition Self-supervised learning Masked autoencoders

来源：评论

学校读者我要写书评

暂无评论

Effective Probabilistic Neural Networks Model for Model-Based Reinforcement Learning USV

引用

IEEE TRANSACTIONS ON AUTOMATION SCIENCE AND ENGINEERING 2025年 22卷 11625-11641页

作者： Huang, Wenjun Cui, Yunduan Li, Huiyun Wu, Xinyu Univ Chinese Acad Sci Sch Artificial Intelligence Beijing 101408 Peoples R China Chinese Acad Sci Shenzhen Inst Adv Technol Shenzhen 518055 Guangdong Peoples R China Chinese Acad Sci Shenzhen Inst Adv Technol CAS Key Lab Human Machine Intelligence Synergy Sys Shenzhen 518055 Guangdong Peoples R China Chinese Acad Sci Shenzhen Inst Adv Technol Guangdong Hong Kong Macao Joint Lab Human Machine Shenzhen 518055 Guangdong Peoples R China

Gaussian process (GP) offers a robust solution for modeling the dynamics of unmanned surface vehicles (USV) in model-based reinforcement learning (MBRL). However, the rapidly increasing computational complexity with a large sample capacity of GP limits its application in complex scenarios that require substantial samples to cover the state space. In this article, a novel probabilistic MBRL approach, probabilistic neural networks model predictive control (PNMPC) is proposed to tackle this issue. With an iterative learning framework, PNMPC properly models the USV dynamics using neural networks from a probabilistic perspective to avoid the computational complexity associated with sample capacity. Employing this model to effectively propagate system uncertainties, a model predictive control (MPC) policy is developed to robustly control the USV against external disturbances. Evaluated by position-keeping and multiple targets-tracking scenarios on a real USV data-driven simulation, the proposed method consistently demonstrates its significant superiority in both model accuracy and control performance compared to not only GP model-based approaches but also the probabilistic neural networks-based MBRL baselines, across various scales of external disturbances.

关键词： Computational modeling Probabilistic logic Neural networks Uncertainty Predictive models Vehicle dynamics Training Stability analysis Computational complexity Adaptation models Reinforcement learning (RL) neural networks predictive models marine vehicles

来源：评论

学校读者我要写书评

暂无评论

RL-Based USV Path Planning under the Marine Multimodal Features Considerations

引用

IEEE Internet of Things Journal 2025年第11期12卷 15274-15287页

作者： Lin, Quanbao Gou, Huaxing Tian, Peidong Zuo, Tian-Yu Zhang, Hanzhong Wang, Xin Sun, Poly Z. H. Shanghai Jiao Tong University School of Aeronautics and Astronautics Shanghai200240 China Southeast University School of Computer Science and Engineering Nanjing211189 China Shanghai University School of Microelectronics Shanghai200444 China Chinese Academy of Sciences CAS Key Laboratory of Human-Machine Intelligence-Synergy Systems Shenzhen Institute of Advanced Technology Guangdong Shenzhen518055 China Guangdong-Hong Kong-Macao Joint Laboratory of Human-Machine Intelligence-Synergy Systems Guangdong Shenzhen518055 China East China Normal University School of Psychology and Cognitive Science Shanghai200062 China Shanghai Jiao Tong University Department of Industrial Engineering Shanghai200240 China

Path planning is an important step in ensuring the safety of Unmanned Surface Vehicle (USV) navigation and executing missions quickly and efficiently. However, current USV path planning methods lack comprehensive consideration of electronic nautical charts and meteorological data, resulting in planned paths being unable to fully utilize marine environmental conditions, which may easily lead to collisions and long navigation times. Based on the above considerations, our study designs a USV path planning system that comprehensively considers the multimodal information from electronic nautical charts and meteorological data. The system consists of three parts: (1) image processing module, (2) meteorological analysis module, and (3) path planning module. In detail, the image processing module obtains the geographical feature information from the electronic chart and constructs a static obstacle environment. The meteorological analysis module obtains the meteorological feature information from meteorological data and constructs a dynamic meteorological vector field environment. The path planning module introduces a designed double deep Q-Network (DQN) structure, a multivariate weighted Dueling network, and a priority sampling mechanism to enhance the DQN algorithm for promising performance in USV path planning. Extensive experiments illustrate the superior performance of the proposed fusion DQN algorithm. Furthermore, the feasibility of the entire path planning system is confirmed. © 2014 IEEE.

关键词： Unmanned surface vehicles

来源：评论

学校读者我要写书评

暂无评论

Robust and Regular Micronano Binary Texture on the Complex Curved Surface for Enhanced Reendothelialization and Antithrombotic Performance

引用

ACS NANO 2025年第3期19卷 3676-3693页

作者： Zhang, Jing Yu, Wenyuan Li, Guoqiang Li, Guiling Chen, Baolan Wang, Luwen Yu, Yang Liu, Zhiyuan Li, Donghai Shandong Univ Adv Med Res Inst Cheeloo Coll Med Jinan 250012 Peoples R China Capital Med Univ Beijing Anzhen Hosp Dept Cardiac Surg Beijing 100029 Peoples R China Chinese Acad Sci CAS Key Lab Human Machine Intelligence Synergy Sys Shenzhen Inst Adv Technol Shenzhen 518055 Peoples R China Guangdong Hong Kong Macao Joint Lab Human Machine Hong Kong 518055 Peoples R China

Blood-contacting medical devices can easily trigger immune responses, leading to thrombosis and hyperblastosis. Constructing microtexture that provides efficient antithrombotic and rapid reendothelialization performance on complex curved surfaces remains a pressing challenge. In this work, we present a robust and regular micronano binary texture on the titanium surface, characterized by exceptional mechanical strength and precisely controlled wettability to achieve excellent hemocompatibility. Systematic in vitro and in vivo investigations confirmed that the micronano binary texture with superhydrophilic modification effectively suppressed the adhesion and activation of plasma proteins and blood cells, thereby mitigating the subsequent coagulation cascade and thrombosis. Meanwhile, the modified surface significantly upregulated the gene expression involving cell-matrix adhesion, growth factor synthesis, and calcium-mediated cytoskeleton remodeling and then accelerated the formation of a healthy and stable endothelial cell layer. This enhancement of re-endothelialization was not observed with pure titanium and superhydrophobic surfaces. Hence, superhydrophilic micronano binary texture not only significantly inhibits thrombosis but also selectively enhances the integrity and viability of the endothelial cell layer, making it a promising strategy for improving the long-term anticoagulation performance of vascular implants.

关键词： micronano binary texture wettability regulation crystalline structure complex curved surface reendothelialization in vivo hemocompatibility

来源：评论

学校读者我要写书评

暂无评论

Dual-interface induced bandgap regulated water-based flexible interactive generator

引用

CHEMICAL ENGINEERING JOURNAL 2025年 511卷

作者： Zhou, Ye Wu, Pengcheng Zhu, Yuchang Yelemulati, Huoerhute Zhu, Shanshan Wu, Keliang Meng, Guihua Wang, Hao Hassan, Fathy Sherif, Mohsen Li, Guanglin Liu, Zhiyong Greish, Yaser E. Tai, Yanlong Chinese Acad Sci Shenzhen Inst Adv Technol Key Lab Human Machine Intelligence Synergy Syst Shenzhen 518055 Peoples R China Shihezi Univ Sch Chem & Chem Engn Shihezi 832003 Xinjiang Peoples R China UAE Univ Natl Water & Energy Ctr Dept Chem Dept Mech Engn Al Ain 15551 U Arab Emirates

Water-based flexible energy electronics are becoming popular for urgent growing demands of the mixed reality interactive technology and the global energy policy of carbon neutrality. However, traditional water-based energy electronics normally show strong selectivity to certain water physical morphologies. To fabricate water- based interactive electronics with high efficiency and can be applied in wider scenarios, in this study, a dual- interface capacitance theory, with Interface-1: electron-charge layer (ECL) and interface-2: ion- charge layer (ICL), was proposed and verified by electrochemical impedance spectroscopy (EIS) analysis and COMSOL simulation. And accordingly, a charge storage material of bandgap regulated polymethyl methacrylate (TPMMA) was developed to fabricate a superior water-based flexible interactive generator, which could present a stable surface potential of 429 V for 5000 h at least, and a sliding potential response of saline water droplet up to 3.57 V, compared with 0.1 V for fluoropolymer as reported. Moreover, the practicality of fluorine-free T-PMMA was explored to harvest the water based mechanical energy, including the volatilization energy of water molecule as test paper, the gravitational potential energy of water droplet as artificial leaf, the mechanical kinetic energy of water flow as vascular disease monitor, respectively.

关键词： Water Energy Harvesting Dual-interface Theory Bandgap Regulation Fluorine-Free Electret Membrane Flexible Interactive Electronics

来源：评论

学校读者我要写书评

暂无评论

Collective Electromagnetically Induced Transparency-Like Slow Light Effect Enhanced by Flatband in All-Dielectric Kagome Metasurfaces

引用

PHYSICA STATUS SOLIDI-RAPID RESEARCH LETTERS 2025年

作者： Fan, Qingxi Huang, Rixing Han, Yingying Luo, Xiaoqing Li, Guangyuan Chinese Acad Sci Shenzhen Inst Adv Technol CAS Key Lab Human Machine Intelligence Synergy Sys Shenzhen 518055 Peoples R China Univ Chinese Acad Sci Shenzhen Coll Adv Technol Shenzhen 518055 Peoples R China Guangxi Univ Sch Elect Engn Nanning 530004 Peoples R China Anhui Normal Univ Coll Phys & Elect Informat Wuhu 241000 Peoples R China

Slow light is a fascinating research area that is of fundamental interest and has enabled diverse applications. Electromagnetically induced transparency (EIT), or its analogies, and flat band are two distinct approaches to realize the slow light effect in photonic systems. Here, these two methods are combined and propose a flatband-enhanced collective EIT-like slow-light effect in all-dielectric kagome metasurfaces. The Q-factor, the group index, and the near-field electric-field enhancement can be significantly improved for the collective EIT-like window enhanced by a flatband, compared to the one without such enhancement, are shown. The simulated transmittance spectra showing these EIT-like windows are experimentally demonstrated. Making use of these collective transparency windows, bulk sensitivities up to 405 and 410 nm/RIU can be numerically obtained. Remarkably, by reducing the nanodisk size, an extremely large group index of 16 105 and high near-field intensity enhancement of 8220 can be obtained for such a silicon kagome metasurface. This work is expected to provide a new direction for realizing strong slow light effect in metasurfaces with promising applications in biochemical sensing, nanolasing, and nonlinear optics.

关键词： all-dielectric metasurfaces collective resonance flatbands kagome lattices slow light

来源：评论

学校读者我要写书评

暂无评论

High-Speed Flexible Near-Infrared Organic Photodetectors for Self-Powered Optical Integrated Sensing and Communications

引用

ADVANCED FUNCTIONAL MATERIALS 2025年第2期35卷

作者： Xia, Yingjun Zhang, Jing Guo, Tingting Wang, Hui Geng, Chen Zhu, Yu Han, Ruiman Yang, Yanqing Song, Guangkun Wan, Xiangjian Li, Guanghui Chen, Yongsheng Nankai Univ Coll Chem Ctr Nanoscale Sci & Technol Tianjin 300071 Peoples R China Nankai Univ Inst Polymer Chem Coll Chem Key Lab Funct Polymer Mat Tianjin 300071 Peoples R China Nankai Univ Frontiers Sci Ctr New Organ Matter State Key Lab Elementoorgan Chem Tianjin 300071 Peoples R China Chinese Acad Sci Shenzhen Inst Adv Technol CAS Key Lab Human Machine Intelligence Synergy Sys Shenzhen 518055 Peoples R China Nankai Univ Renewable Energy Convers & Storage Ctr RECAST Tianjin 300071 Peoples R China

Integrated sensing and communications (ISAC) based on radio frequency (RF), merging the sensing and communication functionalities into a single platform, can effectively address the increasing demand for spectrums due to the proliferation of wireless devices. However, the ISAC commonly suffers from spectrum scarcity, high power consumption, and limited sensing capabilities. Here flexible self-powered optical ISAC (O-ISAC) is proposed and fabricated by preparing a high-speed near-infrared organic photodetector (NIR OPD) that exhibits a broad photodetection ranging from 300 to 1100 nm, specific detectivity over 1013 Jones, and -3 dB cutoff frequency over 1 MHz (lambda = 1050 nm) in self-powered mode, ranking as the fastest among self-powered NIR OPDs in the region above 1000 nm. The device is fabricated by incorporating a narrow-bandgap non-fullerene acceptor BTPSV-4Cl into the PCE-10:BTPSV-4F system to finely tune the film morphology and reduce trap states and energetic disorder. The semitransparent and regular NIR OPDs are integrated on one substrate to fabricate flexible O-ISAC, realizing simultaneous accurate vital signs monitoring and high-speed data transmission under ambient, sunlight, and NIR lighting conditions. It is envisioned that the proposed O-ISAC will inspire a new generation of wireless technology to address RF limitations for emerging applications, such as intelligent healthcare, autonomous vehicles, and security. A flexible self-powered O-ISAC is fabricated based on high-speed NIR OPDs using halogenation and multi-component strategies to finely tune the film morphology and reduce trap states and energetic disorder. Consequently, the flexible O-ISAC realizes simultaneous accurate vital signs monitoring and high-speed optical wireless communication under visible, NIR, and sunlight with various intensities. image

关键词： flexible near-infrared photodetectors narrow bandgap non-fullerene acceptors optical integrated sensing and communications (O-ISAC) optical wireless communication organic photodetectors photoplethysmography

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：