检索结果-内蒙古大学图书馆

Transformer-Based Fusion of RGB and Depth Images for Terrain Recognition

Procedia Computer Science 2024年 250卷 22-29页

作者： Hui Chen Chunjie Chen Zhuo Wang Fangliang Yang Heng Pan Xinyu Wu Shenzhen Institute of Advanced Technology Chinese Academy of Sciences Shenzhen 518055 China Shenzhen College of Advanced Technology University of Chinese Academy of Sciences Shenzhen 518055 China Guangdong Provincial Key Lab of Robotics and Intelligent System Shenzhen Institute of Advanced Technology Chinese Academy of Sciences Shenzhen 518055 China Guangdong-Hong Kong-Macao Joint Laboratory of Human–Machine Intelligence-Synergy Systems Shenzhen 518055 China

This research aims to explore key issues in the field of walking robots and exoskeletons. By using a depth camera to record human walking data in different terrains and employing RGB and depth information for image classification, our approach utilizes both early fusion and late fusion methods to explore their impact on terrain classification. This approach achieved fast and accurate performance in recognizing terrain. We captured data for walking upstairs, downstairs, upward ramps, downward ramps and level ground and used the vision transformer for data fusion. This work compares the performance of unimodal RGB model, unimodal deep model, Pre-Fusion model and Post-Fusion model in terms of accuracy, recall and processing frequency, with accuracies of 0.954, 0.934, 0.976, and 0.976, respectively. The results show that the Post-Fusion model slightly outperforms the Pre-Fusion model in terms of accuracy and recall with more reliable classification performance. To understand the attention mechanism better, we've been visualizing the weights of the model, observing how different attention heads may capture different features of the data from various dimensional perspectives. The visualization helps in interpreting the model's focus and how it distinguishes between features of different nature, such as color texture and spatial depth. These findings can inform the ongoing refinement of terrain classification systems for applications in robotics and autonomous navigation.

关键词： Multimodal fusion terrain recognition vision transformer depth camera

来源：评论

学校读者我要写书评

暂无评论

Dynamic Obstacle Avoidance for Magnetic Helical Microrobots Based on Deep Reinforcement Learning

Dynamic Obstacle Avoidance for Magnetic Helical Microrobots ...

引用

IEEE International Conference on Real-time Computing and robotics (RCAR)

作者： Yukang Qiu Yaozhen Hou Haotian Yang Yigao Gao Hen-Wei Huang Qing Shi Qiang Huang Huaping Wang Intelligent Robotics Institute School of Mechatronical Engineering Beijing Institute of Technology Beijing China Laboratory for Translational Engineering Harvard Medical School Cambridge MA USA Beijing Advanced Innovation Center for Intelligent Robots and Systems Beijing Institute of Technology Beijing China Key Laboratory of Biomimetic Robots and Systems Beijing Institute of Technology Ministry of Education Beijing China

ISBN: (数字)9798350372601

ISBN: (纸本)9798350372618

Magnetic helical microrobots hold immense promise in biomedical domains owing to their compact size and efficient propulsion capabilities. However, navigating these microrobots through dynamic and unstructured environments, particularly when encountering numerous dynamic obstacles, remains a formidable challenge. In the study, a control framework based on deep reinforcement learning (DRL) with the objective of guiding a microrobot through dynamic obstacles towards specified target goals is introduced. Initially, we design and fabricate a microdrill capable of propulsion via external magnetic rotating fields produced by our magnetic actuation system. Subsequently, we construct a custom training environment, adhering to the OpenAI gym interface, to serve as the simulator for training purposes. Utilizing the proximal policy optimization algorithm, we conduct training of the navigation policy within this simulator. Simulations and experimental validations conducted in dynamic environments affirms the efficacy of the proposed method.

关键词： Training Navigation Magnetic cores Heuristic algorithms Magnetic domains Propulsion Deep reinforcement learning

来源：评论

学校读者我要写书评

暂无评论

Exoskeleton-Based Lower Limb Rehabilitation Robot Design and Simulation for Children with Cerebral Palsy

Exoskeleton-Based Lower Limb Rehabilitation Robot Design and...

引用

World Rehabilitation Robot Convention (WRRC)

作者： Wujing Cao Shuo Zhang Yupeng Zou Xinqiang Guo Changyu Li Jingxin Wang Worawarit Kobsiriphat Xinyu Wu Guangdong Provincial Key Lab of Robotics and Intelligent System Shenzhen Institute of Advanced Technology Chinese Academy of Sciences Shenzhen China and SIAT-CUHK Joint Laboratory of Robotics and Intelligent Systems. SIAT Branch Shenzhen Institute of Artificial Intelligence and Robotics for Society School of Mechanical Engineering China University of Petroleum East China Jining Zhongke Intelligent Technology Co. Ltd. Department of Rehabilitation Medicine Zhengzhou Central Hospital Affiliated Zhengzhou University Zhengzhou China National Metal and Materials Technology Center Thailand

ISBN: (数字)9798350355123

ISBN: (纸本)9798350355130

Squat gait is the most common abnormal gait in children with cerebral palsy. Gait training with lower limb rehabilitation robots can improve rehabilitation efficiency and reduce manual labor for them. A novel exoskeleton-based lower limb rehabilitation robot for children with cerebral palsy is proposed, considering the squat gait characteristics of children with cerebral palsy. The structure of the lower limb rehabilitation robot, which combines hip and knee with 4 active degrees of freedom exoskeleton and mobile walker, is designed and optimized. Static analysis of the key mechanical structure of the exoskeleton is conducted using ANSYS Workbench, and local structures are optimized to ensure the safety factor of the mechanical structure is greater than 1, guaranteeing reasonable structural strength and stiffness design. Dynamic simulation analysis based on the ADAMS model validates the reliability and dynamic walking effect of the lower limb exoskeleton rehabilitation robot, obtaining the curve relationship between joint angles and motion cycles, thus verifying the effectiveness of the mechanical system motion.

关键词： Knee Training Cerebral palsy Exoskeletons Dynamics Static analysis Assistive robots Safety Finite element analysis Hip

来源：评论

学校读者我要写书评

暂无评论

Different Height Stairs Climbing Pattern Generation for Humanoids

Different Height Stairs Climbing Pattern Generation for Huma...

引用

IEEE International Conference on Real-time Computing and robotics (RCAR)

作者： Ming Yang Dingkui Tian Minjia Xu Feng Li Jianquan Sun Weiwei Shang Xinyu Wu University of Science and Technology of China Hefei P. R. China Guangdong Provincial Key Lab of Robotics and Intelligent System Shenzhen Institute of Advanced Technology Chinese Academy of Sciences Shenzhen China Daqing Vocational College Daqing China SIAT-CUHK Joint Laboratory of Robotics and Intelligent Systems Shenzhen China

Climbing stairs has been an indispensable ability for humanoids. This paper presents a novel and practical problem: climbing stairs of varying heights. In this study, the humanoid is modeled as an inverted pendulum and set the pelvis trajectory on the vertical axis, representing the center of mass (CoM), as a quadratic function. We also use a polynomial to design the ZMP and CoM trajectory on the horizontal plane. To generate the CoM trajectory, the sliding mode controller is utilized to track the reference trajectory. To eliminate chattering of the sliding mode controller, the universal approximation property of the radial basis function (RBF) network is leveraged to approximate the switching law. Finally, simulation experiments are performed in CoppeliaSim, and the biped robot is capable of walking stably and quickly.

关键词：

来源：评论

学校读者我要写书评

暂无评论

6DoF-ICTC: A Deep Image Feature Extraction Network Combining Transformers and CNNs for 6DoF Pose Estimation 12

6DoF-ICTC: A Deep Image Feature Extraction Network Combining...

引用

12th International Conference on CYBER Technology in Automation, Control, and intelligent systems, CYBER 2022

作者： Zhang, Liming Zhou, Xin Zhu, Baixian Wang, Can Wu, Xinyu Guangdong Provincial Key Lab of Robotics and Intelligent System Shenzhen Institutes of Advanced Technology Chinese Academy of Sciences Shenzhen518055 China Cas Key Laboratory of Human-Machine Intelligence-Synergy Systems Shenzhen Institutes of Advanced Technology China University of Science and Technology of China China Department of Mechanical and Automation Engineering The Chinese University of Hong Kong Hong Kong

ISBN: (数字)9781665472678

ISBN: (纸本)9781665472678

In this work, we propose a new image feature extraction network for 6D pose estimation tasks. The majority of existing 6D pose estimation networks are based on the RGBD input, which is the fusion of images and point clouds. They are mainly working on keypoint extraction, point cloud feature extraction, and other tasks, however, image features, on the other hand, mainly rely on CNN networks that are currently advanced. With the development of Vision Transformer in the field of computer vision, the results on ImageNet demonstrate its benefit of extracting global features, and we see new directions that the Transformers can bring to the area of image feature extraction. ICTC is a Transformer paired with a CNN network that fuses the retrieved global features into each pixel and then does local feature extraction. This provides greater robustness to 6D pose estimation In complex situations such as shifting lighting conditions and object occlusion. Our method achieves 96% accuracy using the average ADD metric on the LineMOD dataset, demonstrating our method's effectiveness. © 2022 IEEE.

关键词： Feature extraction

来源：评论

学校读者我要写书评

暂无评论

Sensory-Motor Neurostimulation to Enhance Exosuit Performance

Sensory-Motor Neurostimulation to Enhance Exosuit Performanc...

引用

IEEE International Conference on Rehabilitation robotics (ICORR)

作者： Chiara Basla Lauren Chee Giacomo Valle Andrea Crema Silvestro Micera Robert Riener Stanisa Raspopovic Sensory-Motor Systems Lab Institute of Robotics and Intelligent Systems (IRIS) ETH Zurich Switzerland NeuroEngineering Lab Institute of Robotics and Intelligent Systems (IRIS) ETH Zurich Switzerland Bertarelli Foundation Chair in Translational Neuroengineering EPFL Neuro-X Institute Switzerland The BioRobotics Institute SantAnna School of Advanced Studies Pisa Italy Spinal Cord Injury Center Medical Faculty Balgrist University Hospital University of Zürich Switzerland

Exosuits typically provide limited mechanical support and rely on a user's residual functional ability. However, people with neurological impairments often suffer from both motor and sensory deficits that limit the assistance an exosuit can provide. To overcome these limitations, we developed the REINFORCE system, that complements the mechanical assistance provided by an exosuit, the Myosuit, with (1) functional electrical stimulation to enhance the activities of leg muscles, and (2) transcutaneous electrical nerve stimulation to restore somatosensory information. It consists of a fully portable and highly modular system that can be easily adapted to the level of impairment and specific need of each participant. Technical verification with three healthy participants showed reliable synchronization between all modules of the systems in all phases of walking. Additionally, we tested the system's effectiveness in one participant with multiple sclerosis who walked overground with and without functional electrical stimulation. Results showed a slight increase in self-selected walking speed (approx. 18%) and in the peak hip flexion at late swing (approx. 12%) as well as reduced step-to-step variability of step length and step time when electrical stimulation was provided. Our findings push towards a clinical trial involving more patients to validate the effectiveness of the REINFORCE system on participants' mobility.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Preliminary study of online real-time control system for lower extremity exoskeletons based on EEG and sEMG fusion

Preliminary study of online real-time control system for low...

引用

2022 IEEE International Conference on robotics and Biomimetics, ROBIO 2022

作者： Li, Wenju Shao, Keyong Zhu, Chi Ma, Yue Cao, Wujing Yin, Meng Yang, Lijun Luo, Mingxiang Wu, Xinyu Northeast Petroleum University School of Electrical Engineering Information Daqing China Shenzhen Institute of Advanced Technology Chinese Academy of Sciences Guangdong Provincial Key Lab of Robotics and Intelligent System Guangdong Shenzhen518055 China Maebashi Institute of Technology Department of Systems Life Engineering Maebashi3710132 Japan

ISBN: (纸本)9781665481090

Motion imaging (MI) is widely used in exoskele-ton robot control direction. However, online real-time con-trol of exoskeletons based on biological signals is difficult to achieve. Because the anti-interference of biological signals in real life is poor and affected by individual differences, the control of auxiliary equipment by the realization of human-machine interface is still facing great challenges. In this paper, a human-machine interface (HMI) based on hybrid biological signals is designed and online real-time control of lower extremity exoskeletons is realized. Based on a comprehensive analysis of human walking, the HMI uses the electromyography signals (upper extremities) and EEG signals generated during real exercise as input signals. Different analysis methods are designed for different modal signals. We extracted four features of sEMG signals to distinguish between different movement processes in the human body. For EEG signals, the use of Independent Component Analysis (ICA) increases the robustness of the system. Event-related desynchronization/synchronization (ERS/ERD) and power spectral densities are reused for analysis from the perspective of the interaction of multiple information in the channel, time and frequency domains. The human-machine interface integrates EEG and EMG multimodal biological signals, and the average recognition accuracy of different subjects has reached 93.7%. In order to increase the recognition accuracy of practical applications and reduce the differences between different individuals, a window voting mechanism is designed. Healthy subjects were tested, and the recognition accuracy of the lower limb exoskeleton online real-time assistance system reached 99%, with high accuracy and reliability. © 2022 IEEE.

关键词： Real time control

来源：评论

学校读者我要写书评

暂无评论

TactCapsNet: Tactile capsule network for object hardness recognition

TactCapsNet: Tactile capsule network for object hardness rec...

引用

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

作者： Fang, Senlin Mi, Tingting Zhou, Zhenning Ye, Chaoxiang Liu, Chengliang Wu, Hancheng Yi, Zhengkun Wu, Xinyu Shenzhen Institute of Advanced Technology CAS Key Laboratory of Human-Machine Intelligence-Synergy Systems Shenzhen518055 China Shenzhen Institute of Advanced Technology Chinese Academy of Sciences Guangdong Provincial Key Lab of Robotics and Intelligent System Shenzhen518055 China Shenzhen Key Laboratory of Smart Sensing and Intelligent Systems Shenzhen Institute of Advanced Technology Chinese Academy of Sciences Shenzhen518055 China SIAT Branch Shenzhen Institute of Artificial Intelligence and Robotics for Society Shenzhen518055 China

ISBN: (纸本)9781665436786

Hardness is one of the most essential tactile clues for robots to recognize objects. However, methods for robots to recognize hardness are limited. In this paper, based on the Capsule Network (CapsNet), we propose a novel tactile capsule network (TactCapsNet) for object hardness recognition. Specifically, we collect a tactile dataset on the silicone samples with three different shapes, and the silicone samples of each shape have thirteen hardness levels ranging from 0A (Shore A scale) to 60A at 5A intervals. Furthermore, we construct the tactile image as the input of the CapsNet to make full use of the spatio-temporal information of the tactile hardness dataset. The experimental results prove that the proposed approach achieves higher accuracy and quadratic weighted kappa (QWK) than support vector machine (SVM), long short-term memory (LSTM), convolutional neural network (CNN), and CapsNet. © 2021 IEEE.

关键词： Hardness

来源：评论

学校读者我要写书评

暂无评论

Lateral Walking Gait Phase Detection with Data Optimization for a Hip Exoskeleton in Resistance Exercise

Lateral Walking Gait Phase Detection with Data Optimization ...

引用

IEEE Workshop on advanced robotics and its Social Impacts, ARSO

作者： Wujing Cao Xinbo Zhu Rui Xu Changyu Li Hongliu Yu Mingming Zhang Haoyong Yu Worawarit Kobsiriphat Xinyu Wu Guangdong Provincial Key Lab of Robotics and Intelligent System Shenzhen Institute of Advanced Technology Chinese Academy of Sciences Shenzhen China SIAT Branch Shenzhen Institute of Artificial Intelligence and Robotics for Society and SIAT-CUHK Joint Laboratory of Robotics and Intelligent Systems. College of Automation Engineering Nanjing University of Aeronautics and Astronautics Jining Zhongke Intelligent Technology Co. Ltd. University of Shanghai for Science and Technology Southern University of Science and Technology Department of Biomedical Engineering National University of Singapore. National Metal and Materials Technology Center Thailand

ISBN: (数字)9798350344639

ISBN: (纸本)9798350344646

The accurate detection of lateral walking gait phases is essential for the effective implementation of hip exoskeleton systems in lateral resistance walking exercises. However, limitations in hardware, such as memory and computing power, in the microcontrollers of wearable devices, significantly impact the size and training speed of the lateral walking gait phase detection model, thus affecting the exoskeleton system. This study proposes a data optimization algorithm that utilizes K-means clustering combined with commonly used machine learning algorithms, including Random Forests (RF), Support Vector Machines (SVM), and k-Nearest Neighbors (KNN), to reduce both the training time and size of the model. With the implementation of this algorithm, the training time and model size of RF, SVM, and KNN-based models are reduced by 89.6%, 99.8%, and 97.9%, and 89.6%, 92.7%, and 95.2% respectively. The corresponding gait phase prediction accuracy experiences only a slight decrease of 1.6%, 1.7%, and 2.8% respectively. This method ensures a sufficiently high accuracy in detecting lateral walking gait phases while simultaneously achieving higher efficiency and a smaller model size.

关键词： Legged locomotion Training Support vector machines Solid modeling Machine learning algorithms Accuracy Exoskeletons

来源：评论

学校读者我要写书评

暂无评论

Mathematical Modelling of a Four-Fingered Metamorphic Anthropomorphic Hand on Precision Grasp

Mathematical Modelling of a Four-Fingered Metamorphic Anthro...

引用

ASME/IFToMM International Conference on Reconfigurable Mechanisms and Robots, ReMAR

作者： Tun Wang Yen-Hua Lin Jian S. Dai Department of Mechanical and Electrical Engineering Centre of Advanced Oil and Gas Equipment China University of Petroleum (Beijing) Beijing China Department of Engineering Centre for Robotics Research King's College London London UK Department of Mechanical and Energy Engineering Shenzhen Key Laboratory of Biomimetic Robotics and Intelligent Systems Southern University of Science and Technology Shenzhen China

ISBN: (数字)9798350395969

ISBN: (纸本)9798350395976

For a metamorphic robotic hand with a foldable palm, its finger operation planes change with palm configurations. Although this reconfigurable design can enhance the grasp flexibility, manipulation workspace and application adaptability, the hand's grasp stability cannot be guaranteed when the palm reconfigures. Especially for the four-fingered metamorphic hand performing precision grasps in anthropomorphic postures, improper control of the palm configurations can result in unsuccessful grasps, which may even lead to hand damage. This paper first displays the workspace of the four-fingered anthropomorphic hand for generating precision grasp gestures. Based on it, force-closure/-balance analyses are performed, deriving the effective precision grasp workspace for typical anthropomorphic postures classified by built mathematical definition. Grasp levels are further discussed by introducing an evaluation method which considers the minimum friction coefficient (MFC) as the quality metric. The proposed scheme gives a way to depict the grasp performance of every robotic hand configuration, without considering specific objects that are to be grasped. Dynamic changes in postures' grasp performance with palm configurations are exhibited in the end.

关键词： Measurement Friction Mathematical models Stability analysis Robots Petroleum

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：