检索结果-内蒙古大学图书馆

Obstacle-Avoidance Motion Planning Method of Manipulator Based on Deep Reinforcement Learning

Unmanned systems 2025年

作者： Liu, Yiyang Wu, Defan Fu, Yukai Bai, Hongfei Deng, Chao State Key Laboratory of Robotics Shenyang Institute of Automation Chinese Academy of Sciences Shenyang110016 China Key Laboratory of Networked Control Systems 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 Institute of Advanced Technology Nanjing University of Posts and Telecommunications Nanjing210023 China

In recent years, driven by the rapid advancement of industrial automation and intelligent manufacturing, the demand for manipulators to perform tasks in complex environments has significantly increased. Traditional motion planning methods face challenges of computational complexity and lack of adaptability when handling robots with redundant degrees of freedom. To address these issues, this paper proposes a novel obstacle avoidance motion planning method for manipulators based on deep reinforcement learning (DRL), which is especially important when unmanned systems operate in a GPS-free environment. The primary contributions are twofold: (i) we propose an innovative algorithm, termed PPO-HER, which incorporates hindsight experience replay (HER) into proximal policy optimization (PPO) during training, effectively leveraging failed experiences to enhance policy learning efficiency;(ii) an adaptive action exploration strategy is designed to accelerate training convergence and avoid local optima. Experimental results demonstrate that the PPO-HER algorithm and the adaptive action exploration strategy perform exceptionally well across three environments of varying difficulty, significantly improving network training speed and success rates, shortening planned paths, and enhancing autonomous obstacle avoidance capabilities. These findings provide new ideas for target location and tracking of unmanned systems, and lay the foundation for future learning and data-driven approaches in complex environments. © World Scientific Publishing Company.

关键词： Deep reinforcement learning

来源：评论

学校读者我要写书评

暂无评论

Design and simulation of a hip exoskeleton for lateral walking

Design and simulation of a hip exoskeleton for lateral walki...

引用

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

作者： Zhang, Zhewen Cao, Wujing Chen, Chunjie Yu, Hongliu Wu, Xinyu Meng, Qiaoling Shenzhen Institute of Advanced Technology Chinese Academy of Sciences Guangdong Provincial Key Lab of Robotics and Intelligent System University of Shanghai for Science and Technology Rehabilitation Engineering and Technology Institute CAS Key Laboratory of Human-Machine Intelligence-Synergy Systems Shenzhen Institute of Advanced Technology

ISBN: (纸本)9781665436786

In this article, a rehabilitation exoskeleton assisted by lateral walking is designed to help patients with lower limb movement disorders to solve the problem of lateral walking assistance, so as to achieve the effect of rehabilitation training. Based on the characteristics of lateral movement of human body, the exoskeleton designed in this paper uses linear push rod motor and thigh binding to assist hip abduction and adduction respectively. Based on the research of human transverse movement gait, a reasonable linkage mechanism was designed and kinematics simulation was carried out. The dynamics model of human lower limbs was established and calculated with MATLAB simulation software. the moment of hip joint lateral movement is obtained. Finally, the structural parameters of the driving mechanism were analyzed by dynamic simulation, and the simulation results were compared with the target motion function to verify the rationality of the design. © 2021 IEEE.

关键词： MATLAB

来源：评论

学校读者我要写书评

暂无评论

Tactile grasp stability classification based on graph convolutional networks

Tactile grasp stability classification based on graph convol...

引用

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

作者： Mi, Tingting Que, Dashun Fang, Senlin Zhou, Zhenning Ye, Chaoxiang Liu, Chengliang Yi, Zhengkun Wu, Xinyu School of Information Engineering Wuhan University of Technology Wuhan430070 China Shenzhen Institute of Advanced Technology CAS Key Laboratory of Human-Machine Intelligence-Synergy Systems Shenzhen518055 China Guangdong Provincial Key Lab of Robotics and Intelligent System Shenzhen Institute of Advanced Technology Chinese Academy of Sciences 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

One of the challenges for robots to grasp unknown objects is to predict whether objects will fall at the beginning of grasping. Evaluating robotic grasp state accurately and efficiently is a significant step to address this issue. In this paper, based on the different fusion approaches of multi-sensor tactile signals, we propose two novel methods based on Graph Convolution Network (GCN) for robotic stability classification. Specifically, we propose two deep learning methods including GCN based on data-level fusion (GCN-DF) and GCN based on feature-level fusion (GCN-FF). We explore the optimal parameters for transforming sensor signals into a graph structure. Furthermore, we verify the effectiveness of the proposed methods on the BioTac Grasp Stability (BiGS) dataset. The experimental results prove that the proposed approaches achieve higher classification accuracy than Support Vector Machine (SVM) and Long Short-Term Memory (LSTM). © 2021 IEEE.

关键词： Convolution

来源：评论

学校读者我要写书评

暂无评论

An intelligent speed-suggestion planner for coverage path with multiple constraints

An intelligent speed-suggestion planner for coverage path wi...

引用

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

作者： Ma, Xinbo Jiang, Guolai Peng, Yuanzhe Ma, Tingchen Liu, Chao Ou, Yong-Sheng Shenzhen Institutes of Advanced Technology Chinese Academy of Sciences Key Laboratory of Human-Machine Intelligence-Synergy Systems University of Chinese Academy of Sciences Beijing100049 China Guangdong Provincial Key Laboratory of Robotics and Intelligent System Shenzhen Institutes of Advanced Technology Chinese Academy of Sciences Shenzhen China Guangdong-Hong Kong-Macao Joint Laboratory of Human-Machine Intelligence-Synergy Systems

ISBN: (纸本)9781665436786

In this paper, we consider various constraints to generate a recommended speed profile for the covered path of the mobile robot, and use the recommended speed profile as the control input to complete trajectory tracking. These speeds are affected by the constraints of the robot's kinematics and dynamics model and the constraints of the robot's environment. The speed planner is used in many scenarios of mobile robots, such as cleaning robots, disinfecting robots, etc., which can improve the safety and efficiency of tasks performed by robots. We introduce the commonly used traffic regulations in daily life into the generation of speed profile. In certain sections of highways or urban roads, there will be a road sign indicating the recommended speed of the section. And our paper is to generate the speed of each point for the given path as the recommended speed for the robot to drive for reference. We use a dynamic window method to update the current environment in real-time. And calculate the influence of the environment in the window on the current speed. Then we use a pure pursuit method to track the given path, and we define a safety cost function to express the safety of the trajectory. We demonstrated the efficiency and security of this method by using a simulation platform. © 2021 IEEE.

关键词： Speed

来源：评论

学校读者我要写书评

暂无评论

Dynamic Fluorescent Imaging for Robotic Patch Clamp in vivo: Modeling and Control

Dynamic Fluorescent Imaging for Robotic Patch Clamp in vivo:...

引用

Chinese Control Conference (CCC)

作者： Ruimin Li Jinyu Qiu Minghui Li Mingzhu Sun Xin Zhao Qili Zhao Institute of Robotics and Automatic Information System (IRAIS) Nankai University Tianjin China Tianjin Key Laboratory of Intelligent Robotic (tjKLIR) Nankai University Tianjin China Institute of Intelligence Technology and Robotic Systems Shenzhen Research Institute of Nankai University Shenzhen China Beijing Advanced Innovation Center for Intelligent Robots and Systems Beijing Institute of Technology Beijing China

Fluorescent imaging through fluorescent dye injection is the key technique for lighting the in vivo environment and localizing the target cell in it for patch clamp operation in vivo. At present, manual fluorescent imaging is usually a labor-intensive high-skill requisite work, significantly limiting the application of patch clamp operation in vivo. In this paper, a computational modeling of the diffusion process of the injected fluorescent dye is established using the commercial ANSYS Fluent software. Based on this model, the distribution of the mass fraction, proven to be linearly related to fluorescence intensity, was calculated to simulate the fluorescent field formed by fluorescent dye injection. Further, an ADRC controller was developed to generate feedback-controlled fluorescent fields. With the above model and controller, a mobile fluorescent field, and varying fluorescent field with constant mass fraction at the target position are generated to fulfill the requirement of in vivo environment observation, search and localization of the target cell in patch clamp operation, respectively. The above results prove the feasibility of ADRC-controlled dynamic fluorescent imaging in the patch clamp operation in vivo.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Accurate Prediction of Knee Joint Angles Using a Hybrid CNN-LSTM-Attention Network from Surface Electromyography

Accurate Prediction of Knee Joint Angles Using a Hybrid CNN-...

引用

IEEE International Conference on robotics and Biomimetics

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

ISBN: (数字)9798331509644

ISBN: (纸本)9798331509651

In various fields such as robotics, rehabilitation, biomechanics, human-machine interfaces, and clinical research, human motion prediction has extensive applications. A hybrid networks model based on Convolutional Neural Network-Long Short-Term Memory-Attention (CNN-LSTM-Attention) was employed in this paper to extract features from the high-dimensional space and time series of surface electromyography (sEMG) data for continuous prediction of knee angles. Five able-bodied subjects participated in the study, walking at a self-selected pace while six surface electromyography (sEMG) signals from muscles involved in knee flexion and extension, along with knee angles, were synchronously recorded. The performance was assessed using the root mean square error (RMSE), Pearson Correlation Coefficient $(r)$ , and coefficient of determination $(R^{2})$ between estimated and measured knee joint angles. Compared to the CNN, LSTM, and CNN-LSTM algorithms, the adopted method estimated knee angles more accurately, achieving an average RMSE, $\rho$ , and $R^{2}$ of $3.09\pm 0.90,\ 0.9911 \pm 0.0052$ , and $0.9817\pm 0.0108$ , respectively. As the advance time increased, the accuracy of knee joint angle estimation decreased, particularly when the advance time reached 100 ms, at which point the RMSE is 4.83 degress. In addition, the computing time of the prediction model was less than 1 ms, which was shorter than the advance time 100 ms it could predict, thereby facilitating real-time computation in practical environments. Compared to previous studies, the proposed method excelled at extracting features directly from raw signals, avoiding the need for handcrafted feature extraction. Consequently, the model effectively achieved continuous and accurate predictions of joint motion angles and demonstrated potential to reduce latency in the human-machine interface.

关键词： Knee Accuracy Computational modeling Predictive models Feature extraction Electromyography Convolutional neural networks Wearable devices Long short term memory Robots

来源：评论

学校读者我要写书评

暂无评论

Learning Compliant Assembly Strategy From Demonstration

Learning Compliant Assembly Strategy From Demonstration

引用

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

作者： Sheng Liu Juyi Sheng Yongsheng Ou Southern University of Science and Technology (SUSTech) Guangdong Provincial Key Laboratory of Robotics and Intelligent System Shenzhen Institute of Advanced Technology Chinese Academy of Sciences Shenzhen China Guangdong-Hong Kong-Macao Joint Laboratory of Human-Machine Intelligence-Synergy Systems (#2019B121205007)

Compared with robots, humans can complete the different assembly tasks of parts flexibly and quickly. By teaching robots with human experiences, not only the industrial assembling tasks can be resolved, but also many other robot applications can be realized. The pose adjustment stage is the most critical part of the peg-in-hole assembly process. This paper analyzes the contact force in the pose adjustment stage and calculates the control variables that affect the assembly motion. Based on the Gaussian mixture model (GMM), the nonlinear mapping relationship between the control variables and the state variables during the assembly is established using the human demonstration data, and the parameters of the model are solved by the expectation maximization (EM) algorithm, and thus the human-like compliant assembly is completed by a robot. In order to verify the effectiveness of the demonstration-learning algorithm, a peg-in-hole assembly experiment was carried out using KUKA manipulator. Finally, the experiment shows that the proposed learning-based method not only improves the efficiency of the robot peg-in-hole assembly but also makes the manipulator have a satisfied adaptive ability to the complex environments in the assembly process.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Admittance Control based Hemiplegic Exoskeleton Motion Control

Admittance Control based Hemiplegic Exoskeleton Motion Contr...

引用

2022 IEEE International Conference on Real-Time Computing and robotics, RCAR 2022

作者： He, Bailin Zhou, Xin Huai, Yaping Liu, Jiaqing Li, Pengbo Wang, Can Wu, Xinyu Guangdong Provincial Key Laboratory of Robotics and Intelligent System Shenzhen Institutes of Advanced Technology Chinese Academy of Sciences Shenzhen518055 China Shenzhen Institutes of Advanced Technology Chinese Academy of Sciences CAS Key Laboratory of Human-Machine Intelligence-Synergy Systems Shenzhen518055 China University of Chinese Academy of Sciences Beijing100049 China Department of Mechanical and Automation Engineering Chinese University of Hong Kong Hong Kong Shenzhen College of Advanced Technology University of Chinese Academy of Sciences Shenzhen518055 China Guangdong-Hong Kong-Macao Joint Laboratory of Human-Machine Intelligence-Synergy Systems Shenzhen Institutes of Advanced Technology Chinese Academy of Sciences Shenzhen518055 China Department of Rehabilitation Medicine Shenzhen Longhua District Central Hospital Shenzhen518000 China

ISBN: (数字)9781665469838

ISBN: (纸本)9781665469838

In this paper, a method based on admittance control is developed to improve patients' comfort wearing an exoskeleton robot for rehabilitation training. In our previous work, we designed a modular and convenient hemiplegic exoskeleton robot. The mechanical structure will cause stiffness when it comes to contact with the human body in practical applications. We introduced admittance control to improve the flexibility of the exoskeleton robot and improve the subjects' comfort. This paper considers coupling of model between the exoskeleton and the human body as the human-computer interaction force of the system at the main human-computer interaction point. Measuring the interaction force between exoskeleton members and limbs, the work uses the virtual admittance model to correct the system's inherent characteristics and studies the control method's stability with Lyapunov's theory. The experiment collected the subject movement information which was converted the movement information of the exoskeleton. Human-computer interaction forces in the thigh and calf were measured while the exoskeleton assisted the subject's movement. The experimental results show that adding the admittance controller reduces the human-computer interaction force from 5g to 2g and 6g to 3g, respectively. © 2022 IEEE.

关键词： Human computer interaction

来源：评论

学校读者我要写书评

暂无评论

Performance benchmarking of YOLO architectures for vehicle license plate detection from real-time videos captured by a mobile robot 16

Performance benchmarking of YOLO architectures for vehicle l...

引用

16th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications, VISIGRAPP 2021

作者： Ismail, Amir Mehri, Maroua Sahbani, Anis Ben Amara, Najoua Essoukri Université de Sousse Ecole Nationale d’Ingénieurs de Sousse LATIS-Laboratory of Advanced Technology and Intelligent Systems Sousse4023 Tunisia Enova Robotics Novation City Technopôle de Sousse Sousse4000 Tunisia

ISBN: (纸本)9789897584886

In this paper, we address the issue of vehicle license plate (LP) detection for a mobile robotic application. Specifically, we tackle the dynamic scenario of a robot in the physical world interacting based on its cameras. The robot is dedicated essentially to patrol and secure unconstrained environments. Counter to the most recent works of LP detection which assume controlled deploying scenario, the mobile platform requires a more robust system that is suitable for various complex scenarios. To contribute to this purpose, we propose an end-to-end detection module capable of localizing LP either in images or in live-streaming videos. The proposed system is based on deep learning based detectors, particularly the most recent YOLOv4-tiny one. To evaluate the proposed system, we introduce the first-ever public Tunisian dataset, called PGTLP, for LP detection that contains 3,000 annotated images. This dataset was gathered using the security robot during its patrolling and surveillance of parking stations and high-risk areas. For the detection, a comparative study for the different YOLO variants has been carried out in order to select the best detector. Our experiments are performed on the PGTLP images and following the same experimental protocol. Among the selected models, YOLOv4-tiny reveals the best compromise between detection performance and complexity. Further experiments that have been conducted using the AOLP benchmark dataset point out that the proposed system has satisfying results. Copyright © 2021 by SCITEPRESS – Science and Technology Publications, Lda. All rights reserved.

关键词： Deep neural networks

来源：评论

学校读者我要写书评

暂无评论

A motion planning method based on HRL for autonomous exoskeleton

A motion planning method based on HRL for autonomous exoskel...

引用

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

作者： Dong, Yao He, Yong Su, Zhilong Ni, Jiangpeng Feng, Wei Wu, Xinyu Shenzhen Institute of Advanced Technology Chinese Academy of Sciences Guangdong Provincial Key Laboratory of Robotics and Intelligent System Shenzhen518055 China SIAT Branch Shenzhen Institute of Artificial Intelligence and Robotics for Society Shenzhen College of Advanced Technology University of Chinese Academy of Sciences Shenzhen518055 China CAS Key Laboratory of Human-Machine Intelligence-Synergy Systems Shenzhen Institute of Advanced Technology Chinese Academy of Sciences Shenzhen518055 China

ISBN: (纸本)9781665436786

With the increase of the world's population and the serious aging phenomenon, more and more people with mobility impairments need technology to assist them in their daily lives. Autonomous Exoskeleton are one of the important technologies. The existing gait generation algorithm are most based on simplified mathematical model calculations, which have the problem of poor accuracy and unfriendliness to the wearer. This research aims to propose a path planning and gait generation policies (HPPO) based on hierarchical reinforcement learning (HRL), so that the wearer can safely avoid obstacles and maintain dynamic balance when using the autonomous exoskeleton controlled by this policy to walk. The high-level controller (HL) is responsible for path planning, it can quickly plan a reasonable exoskeleton footholds in the indoor environment. The low-level controller (LL) can imitate a given reference trajectory, learn a walking policy suitable for the exoskeleton, and it can follow the planned footholds, so as to achieve the purpose of avoiding obstacles for the wearer. The algorithm is verified in simulation environment. © 2021 IEEE.

关键词： Motion planning

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：