检索结果-内蒙古大学图书馆

arXiv 2024年

作者： Zhang, Yuanwen Xiong, Jingfeng Xian, Haolan Chen, Chuheng Chen, Xinxing Fu, Chenglong Leng, Yuquan The Shenzhen Key Laboratory of Biomimetic Robotics and Intelligent Systems Guangdong Provincial Key Laboratory of Human-Augmentation and Rehabilitation Robotics in Universities Southern University of Science and Technology Shenzhen518055 China The Department of Mechanical and Energy Engineering Southern University of Science and Technology Shenzhen518055 China

The hip joint moment during walking is a crucial basis for hip exoskeleton control. Compared to generating assistive torque profiles based on gait estimation, estimating hip joint moment directly using hip joint angles offers advantages such as simplified sensing and adaptability to variable walking speeds. Existing methods that directly estimate moment from hip joint angles are mainly used for offline biomechanical estimation. However, they suffer from long computation time and lack of personalization, rendering them unsuitable for personalized control of hip exoskeletons. To address these challenges, this paper proposes a fast hip joint moment estimation method based on generalized moment features (GMF). The method first employs a GMF generator to learn a feature representation of joint moment, namely the proposed GMF, which is independent of individual differences. Subsequently, a GRU-based neural network with fast computational performance is trained to learn the mapping from the joint kinematics to the GMF. Finally, the predicted GMF is decoded into the joint moment with a GMF decoder. The joint estimation model is trained and tested on a dataset comprising 20 subjects under 28 walking speed conditions. Results show that the proposed method achieves a root mean square error of 0.1180 ± 0.0021 Nm/kg for subjects in test dataset, and the computation time per estimation using the employed GRU-based estimator is 1.3420 ± 0.0031 ms, significantly faster than mainstream neural network architectures, while maintaining comparable network accuracy. These promising results demonstrate that the proposed method enhances the accuracy and computational speed of joint moment estimation neural networks, with potential for guiding exoskeleton control. Copyright © 2024, The Authors. All rights reserved.

关键词： Exoskeleton (robotics)

来源：评论

学校读者我要写书评

暂无评论

A Fast-trained and Generalized Spiking Neural Network for Robotic Gas Source Localization

A Fast-trained and Generalized Spiking Neural Network for Ro...

引用

International Conference on Advanced robotics and Mechatronics (ICARM)

作者： Chiyang Liu Xinxing Chen Jian Huang Hubei Key Laboratory of Brain-inspired Intelligent Systems and the Key Laboratory for Image Processing and Intelligent Control School of Artificial Intelligence and Automation Huazhong University of Science and Technology Wuhan China Guangdong Provincial Key Laboratory of Human-Augmentation and Rehabilitation Robotics in Universities Southern University of Science and Technology Shenzhen China

ISBN: (数字)9798350385724

ISBN: (纸本)9798350385731

Mobile robots for Gas Source Localization (GSL) tasks are a safer alternative than human and animal rescuers in hazardous scenarios. Existing research primarily concentrates on rule-based algorithms or conventional artificial neural networks (ANNs). However, these approaches are either inapplicable in cluttered environments or undeployable due to their high energy consumption and the demand for substantial computational resources. This paper introduces the application of energy-efficient spiking neural networks (SNNs) to address robotic GSL tasks. A pipeline is proposed to train SNNs with deep Q learning and a pretrain-finetune paradigm. To facilitate the training process, a small dataset of gas dispersion is generated utilizing openFoam and GADEN, a high-fidelity simulator for gas dispersion. Data from a simplified plume model are leveraged to pretrain an ANN, the activation function of which gradually transitions from a bounded rectified linear unit (bReLU) to a step function. Subsequently, an SNN initialized with the ANN parameters undergoes finetuning on the GADEN-based dataset. The training pipeline significantly reduces training time compared to direct training of SNNs. The trained SNN is validated within the GADEN simulation environment and compared to three different models, demonstrating promising performance and superior generalization despite limited training data.

关键词： Training Location awareness Q-learning Mechatronics Pipelines Training data Spiking neural networks Data models Mobile robots Dispersion

来源：评论

学校读者我要写书评

暂无评论

Design of magnetorheological brake with adjustable interdigitated comb for lower limb prostheses and exoskeletons

Design of magnetorheological brake with adjustable interdigi...

引用

2023 IEEE International Conference on robotics and Biomimetics, ROBIO 2023

作者： Shi, Qiuyu Zhao, Xuan Liao, Haisu Lin, Yongji Huang, Ruisen Wu, Xinyu Gao, Fei Shenzhen Institute of Advanced Technology Chinese Academy of Sciences Guangdong Provincial Key Laboratory of Robotics and Intelligent System Shenzhen518055 China Guangdong-Hong Kong-Macao Joint Laboratory of Human-Machine Intelligence-Synergy Systems Shenzhen518055 China The Chinese University of Hong Kong Department of Mechanical and Automation Engineering Shatin 999077 Hong Kong

ISBN: (纸本)9798350325706

Magnetorheological (MR) fluids-based devices such as MR dampers and brakes can output a continuously controllable damping force via modulating the magnitude of the applied magnetic field. Owing to low energy consumption and fast response, recently vast MR dampers and brakes are widely studied and explored for lower limb prostheses and exoskeletons. This paper proposed a novel MR brake with adjustable interdigitated comb for lower limb prostheses. Through actively adjusting the area the of the comb engagement, the maximum viscous and field-induced damping torque can be increased or decreased. Because of that, the dynamic adjustment interval of the output damping torque can be increased to meet the function requirements of the prosthetic knee in different walking modes. Magnetic analysis via ANSYS is performed to check the magnetic density of the MR fluid under different excitations. Mathematical models of the damping torque are built. Finally, a prototype of the novel MR brake is fabricated. © 2023 IEEE.

关键词：

来源：评论

学校读者我要写书评

暂无评论

SAF: Local Shape-aware Face-based Garment Collision Handling via Neural SDFs

SAF: Local Shape-aware Face-based Garment Collision Handling...

引用

2025 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2025

作者： Tang, Minzhe Yuan, Ruisheng Kou, Dongliang Sun, Mingyang Zhang, Lihua Academy for Engineering and Technology Fudan University Shanghai China Institute of Metaverse & Intelligent Medicine Fudan University Shanghai China Jilin Provincial Key Laboratory of Intelligence Science and Engineering Changchun China Engineering Research Center of AI and Robotics Ministry of Education Shanghai China AI and Unmanned Systems Engineering Research Center of Jilin Province Changchun China

ISBN: (纸本)9798350368741

Learning-based garment prediction presents an appealing alternative to physics-based methods owing to its high efficiency. However, the predicted garments can exhibit noticeable penetrations into the body. Many collision-handling methods operate in the point domain, which has an inherent limitation in addressing the penetration issues with the mesh faces. In light of this, we propose a local Shape-Aware, Face-based collision handling approach (SAF) that can be applied to garment prediction networks to achieve real-time collision handling. Considering the nature of garments, we design a compact formulation to model the continuity of the garment surface, and utilize neural Signed Distance Fields (SDFs) to accomplish penetration resolution. Recognizing the significant impact of the body's local shape on collision handling, we further propose using the angles between SDF gradients to characterize the sharpness of the body. Our approach can compensate for the inaccuracy of neural SDFs and preserve local smoothness and details. Extensive experiments demonstrate the outstanding performance and generalizability of our method. © 2025 IEEE.

关键词： collision handling computer graphics deep learning garment simulation

来源：评论

学校读者我要写书评

暂无评论

Trajectory Planning for Formation Variation of UGVs in Cluttered Environment

Trajectory Planning for Formation Variation of UGVs in Clutt...

引用

International Conference on intelligent Transportation

作者： Jie Cheng Ruishuang Chen Zhihui Liang Shuan Ding Zaiyue Yang Shenzhen Key Laboratory of Control The-ory and Intelligent Systems School of System Design and Intelligent Manufacturing Guangdong Provincial Key Laboratory of Human-Augmentation and Rehabilitation Robotics in Universities Southern University of Science and Technology (SUSTech) Shenzhen China

The problem of trajectory planning for formation variation of unmanned ground vehicles (UGVs) is challenging due to the coupled non-collision constraints that apply to multiple vehicles. To address this issue, an innovative frame-work based on the alternating direction method of multipliers (ADMM) is proposed in this paper. The framework decouples complex constraints into two categories: individual constraints and mutual constraints. The former category involves dynamic, formation, and static obstacle constraints, which can be solved in parallel by each vehicle. Convex feasible set (CFS) algorithm is employed in this paper to simplify the static obstacle constraints. The latter category deals with collision avoidance among vehicles during the variation process and can be solved by nonlinear solvers. The convergence of the ADMM for the optimization problem is proved in this paper. Compared with solving the entire problem directly, the proposed framework reduces computation time and improves the quality of solution.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Environmental Recognition and Multimode Continuous-Phase Control for a Powered Transfemoral Prosthesis

Environmental Recognition and Multimode Continuous-Phase Con...

引用

IEEE International Conference on Development and Learning, ICDL

作者： Shucong Yin Xinxing Chen Teng Ma Yuxuan Wang Yixuan Guo Yuquan Leng Chenglong Fu Shenzhen Key Laboratory of Biomimetic Robotics and Intelligent Systems Guangdong Provincial Key Laboratory of Human-Augmentation and Rehabilitation Robotics in Universities Southern University of Science and Technology Shenzhen China Department of Biomedical Engineering National University of Singapore.

The continuous-phase-based control method is currently being developed for powered transfemoral prostheses. However, most of the research concentrates on single locomotion mode. The multimode continuous-phase control method for powered transfemoral prostheses is still an open question. To solve the existing problem, this paper proposes a multimode continuous-phase control method for powered transfemoral prostheses. The prosthesis first needs to recognize the surrounding environment and obtain the corresponding environmental features. Then, the gait phase is estimated by phase variable which represents the mapping relationship between thigh angle and gait phase. Finally, the desired trajectories of the knee and ankle joints on various locomotion modes are constructed as functions of the gait phase and environmental features. These functions are called virtual constraints. The proposed phase variable and virtual constraints are evaluated on a public dataset through offline analysis. The root mean squared errors (RMSEs) of the phase variable and virtual constraints in all locomotion modes are less than 3 % and 4 %, respectively. It proves that the proposed method can precisely predict the gait phase and joints angle for powered prostheses in all locomotion modes.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Softening Nonlinear-Stiffness Elastic Mechanism with Continuous Adjustability for Human-Robot Interaction Force Control

SSRN

引用

SSRN 2024年

作者： Tu, Zhixin Liu, Haifeng Jiang, Yihao Ye, Tao Qian, Yuepeng Leng, Yuquan Dai, Jiansheng Fu, Chenglong Shenzhen Key Laboratory of Biomimetic Robotics and Intelligent Systems Guangdong Provincial Key Laboratory of Human-Augmentation and Rehabilitation Robotics in Universities Southern University of Science and Technology Guangdong Shenzhen518055 China Centre for Robotics Research King’s College London LondonWC2R 2LS United Kingdom

Human-robot interaction of human augmentation robots presents a considerable challenge in achieving accurate and robust interaction force control. This paper proposes a novel adjustable nonlinear elastic mechanism with softening stiffness to address this challenge. The diamond elastic mechanism achieves softening stiffness behavior through a nonlinear mapping relationship between the lengths of the diamond diagonals. This unique stiffness profile, featuring high stiffness for low output and low stiffness for high output, strikes a balance between the output force range and force resolution. Moreover, the adjustable stiffness profile is realized by utilizing two antagonistic linear springs, enabling optimal stiffness matching for different output force ranges. Bench tests were conducted to validate the stiffness modeling and evaluate the force tracking and interaction performance of the developed elastic mechanism. Experimental results demonstrate the capability of the diamond elastic mechanism to achieve precise force control and good collision safety in human-robot interaction. Specifically, the root-mean-square error of step response and sinusoidal tracking are all lower than 1% (minimum 0.2704±0.0399 N) and 2% (minimum 0.9254±0.1349 N) of the reference forces, respectively. © 2024, The Authors. All rights reserved.

关键词： Force control

来源：评论

学校读者我要写书评

暂无评论

Learning to Synthesize and Remove Rain Unsupervisedly 18th

Learning to Synthesize and Remove Rain Unsupervisedly

引用

18th Pacific Rim International Conference on Artificial Intelligence, PRICAI 2021

作者： Qi, Yinhe Pan, Meng Jin, Zhi School of Intelligent Systems Engineering Sun Yat-sen University Shenzhen China Guangdong Provincial Key Laboratory of Fire Science and Technology Guangzhou510006 China Guangdong Provincial Key Laboratory of Robotics and Digital Intelligent Manufacturing Technology Guangzhou510535 China

ISBN: (纸本)9783030893699

Most existing single image deraining networks are trained in a supervised way, which relies on paired images including one clean image and one rain image. In most cases, the rain images are synthesized from the clean ones manually to obtain sufficient paired images. However, not only huge time costs but expert knowledge are needed to ensure the synthesized images are realistic enough. In addition, the superior performance of these deraining networks trained on manually synthesized rain images is hard to be maintained when testing on real rain images. To address these issues, we propose a scene adaptive asymmetric CycleGAN (SAA-CycleGAN) which transfers clean images to their rainy counterparts automatically so that adequate realistic rain images can be obtained for training deraining networks in a supervised way. Moreover, SAA-CycleGAN can both remove rain from rainy images and synthesize rain on clean images benefiting from the cycle consistency strategy. Since the information is not symmetric during the rain synthesis process and the deraining process, the generators are designed with different architecture accordingly for these two processes. Comprehensive experiments show that the SAA-CycleGAN is able to synthesize more lifelike rain images and achieve similar deraining performance compared with the state-of-the-art deraining methods. © 2021, Springer Nature Switzerland AG.

关键词： Rain

来源：评论

学校读者我要写书评

暂无评论

SWhegPro3: A Three-Impeller Wheel-Leg Transformable Robot with Variable Robust Adaptability to Stair Dimensions

SWhegPro3: A Three-Impeller Wheel-Leg Transformable Robot wi...

引用

IEEE International Conference on robotics and Biomimetics

作者： Haoran Wang Siyuan Wang Cunxi Dai Zhenzhong Jia Department of Mechanical and Energy Engineering Shenzhen Key Laboratory of Biomimetic Robotics and Intelligent Systems Southern University of Science and Technology (SUSTech) Shenzhen China Guangdong Provincial Key Laboratory of Human-Augmentation and Rehabilitation Robotics in Universities SUSTech Shenzhen China

This paper reports the design innovation and implementation of a novel wheel-leg transformable robot named SWhegPro3. The robot uses a three-impeller design to significantly improve the overall stair-climbing performance. Unlike other impeller-based designs that use gear sets to unfold the impellers, this robot uses self-locking electric push rods for mode switching, thereby greatly improving its load capacity. We also analyze the stair climbing operation and derive a model to calculate the optimal configuration of the wheel-leg module and the corresponding displacement of the push rod, in order to better adapt to different sized stairs. We conduct both simulation and field experiments to validate our proposed design and control strategy.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Modeling and Experimental Characterization of Propeller-driven Capsule Endoscope Robot for Gastrointestinal Minimally Invasive Diagnosis

Modeling and Experimental Characterization of Propeller-driv...

引用

IEEE International Conference on Cyborg and Bionic systems (CBS)

作者： Weihao Wang Yi Zhang Chengzhi Hu Department of Mechanical and Energy Engineering Shenzhen Key Laboratory of Biomimetic Robotics and Intelligent Systems Southern University of Science and Technology Shenzhen China Guangdong Provincial Key Laboratory of Human-Augmentation and Rehabilitation Robotics in Universities Southern University of Science and Technology Shenzhen China

Wireless capsule endoscopy robots are miniaturized robots that integrate camera, wireless communication, microcontroller, power supply, and other modules. In this paper, we study the modeling and experimental characterization of a propeller-driven capsule endoscopy robot (PCER). The profile of the propeller is challenging to design without detailed drawings because of the complex curved surfaces. The propeller design is achieved in the MIT open-source project OpenProp. The modeling of the propeller and the PCER shell are completed. We used the FLUENT fluid simulation toolkit to evaluate the hydrodynamic performance of the propeller and PCER shell. The thrust generated by the propeller is measured by experiment, and the resistance of PCER shell is measured by experiment following the fluid similarity law. The error between the simulation and the experiment results is within the allowable range, which proves that the hydrodynamic performance parameters of the micro-propeller and the PCER shell calculated by the simulation are reliable, and it provides a new method for the design and optimization of the PCER.

关键词： Wireless communication Solid modeling Fluids Spirals Propellers Endoscopes Propulsion

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：