This paper delves into the intricate field of inverse kinematics in the context of the Chinese Fusion Engineering Test Reactor (CFETR) and its specialized robot, the multi-purpose overload robot (CMOR). Nuclear fusion, a cornerstone for sustainable energy, presents immense potential with benefits like high energy output and minimal environmental impact. The goal of the CFETR is to transform nuclear fusion energy from theory to practical solutions to solve the global energy crisis. The CMOR, with nine degrees of freedom, is pivotal in maintaining in-vessel components damaged by intense radioactive exposure. This study introduces a novel approach to solve the inverse kinematics of the CMOR based on the CFETR's work environment. It involves dividing the working space, identifying target interval based on coordinates, and employing intelligent algorithms for precise and efficient kinematic resolution. Six algorithms were evaluated, highlighting the proposed approach’s superior computational speed and accuracy, crucial for real-time monitoring and control in CFETR operations. This work presents a significant advancement in redundant robotic kinematics in nuclear fusion environment.

This study concentrates on the engineering design and analysis of the root joints for the CFETR Multi-Purpose Overload Robot (CMOR), a crucial component in the China Fusion Engineering Test Reactor (CFETR). The CFETR tokamak is an engineering test reactor in China for magnet confinement fusion research. CMOR is a 7-DOF manipulator characterized by heavy-load capacity, high precision, and compact structures, designed for internal maintenance inside the CFETR. This presents significant challenges in its design. The design of the root joints is significant for the CMOR. This paper provides an in-depth exploration of the engineering design and analysis of the root joints. It offers a thorough explanation of the driving systems' design, including an integrated rescue module to address potential manipulator faults. Additionally, this paper delves into the conceptual design and manufacturing schemes of the corresponding shells, which are validated under various loading conditions—including regular and earthquake scenarios—through finite element analysis. The assembled root joints' internal structure and cable routing scheme are also discussed in-depth. This research provides valuable insights for future joint designs of the CMOR heavy load manipulator, advancing maintenance operations for fusion reactors.

Defect recognition of flat metals is paramount for ensuring quality control during the production process. However, the diverse origins of metal surface damage, ranging from mechanical impacts to chemical corrosion, and the resulting varied morphology and scale of surface defects, particularly numerous microdefects and elongated defects with high aspect ratios, complicate defect recognition. Existing methods fail to select the most beneficial features during extraction and commonly lose critical feature information during gradient sampling. To overcome these challenges, we propose a lightweight network to optimize feature screening for defect recognition. First, we propose a deformable context-guided block that employs deformable convolution to dynamically adapt the perception of the spatial context, providing precise guidance of relevant semantic information in complex surface textures. Second, we develop a content-aware feature compression block that implements adaptive weighting of features, which significantly reduces information loss during the downsampling stage. Finally, we introduce an intra-scale feature interaction transformer block, which optimizes high-order semantic features to enhance the accuracy and reliability of defect detection. Experimental validation on the NEU-DET, APS-DET, and GC10-DET datasets demonstrated significant improvements in the detection accuracy and parameter efficiency, confirming the proposed method's robust generalizability.

Recently, neural network-based methods for multi-label textile fiber recognition have achieved considerable success. However, a significant limitation of most current approaches is their disregard for the valuable dependencies that exist among different fiber categories. The universal multi-label image recognition methods that consider label relationships often fall short when applied to the challenge posed by the mixture of fibers in textile images. And these relationship modeling manners have not yet been used in the field of multi-label textile fiber recognition. In this work, a graph relationship-driven method is proposed for the recognition of multi-label textile fibers. Based on the graph attention network, the proposed method introduces global relation graph, label coded mapping, and label compensation to equip the feature learning backbone with the relationship modeling ability. First, feature learning backbone extracts the semantic features from images, and global relation graph mines shallow representations of global label relationships from label embeddings. Then, label coded mapping combines these semantic features and global features to model joint relationships. The obtained joint representations constitute the label code, which is used to map the fiber class indices. Finally, label compensation further extracts the deep representations of global label relationships and utilizes them to compensate the fiber class indices to obtain final prediction scores. By employing experiments on a textile fiber dataset and a rearranged dataset, the effectiveness and superiority of the proposed method are validated. Further experiments on PASCAL VOC 2007 showcase the potential applications of the proposed method beyond textile fiber recognition.

Tokamak-based controlled nuclear fusion is widely acknowledged as the most viable path towards sustainable energy development. China Fusion Engineering Test Reactor (CFETR) fusion teleoperation program utilizes a seven-axis heavy-duty robotic arm to maintain its Tokamak device. This paper suggests employing digital twin technology for the operational maintenance of CFETR’s Multipurpose Overload Robot (CMOR), and discusses the development of a digital twin system according to its maturity level, ranging from simple to complex. Through real-time mapping of the heavy-duty manipulator’s motion planning control system with a virtual physics engine, we verify the feasibility of this digital twin system solution and anticipate future work in this area.

The CFETR multi-purpose overload robot (CMOR) is a key subsystem of the remote handling system of the China fusion engineering test reactor (CFETR). This paper first establishes the kinematic and dynamic models of CMOR and analyzes the working process in the vacuum chamber. Based on the uncertainty of rigid-flexible coupling, a CMOR adaptive robust sliding mode controller (ARSMC) is designed based on the Hamilton-Jacobi equation to enhance the robustness of the control system. In addition, to compensate the influence of non-geometric factors on position accuracy, an error compensation method is designed. Based on the matrix differentiation method, the CMOR coupling parameter errors are decoupled, and then the gridded workspace principle is used to identify the parameter errors and improve the motion control accuracy. Finally, the CMOR rigid-flexible coupling simulation system is established by ADAMS-MATLAB/Simulink to analyze the dynamic control effect of ARSMC. The simulation results show that the CMOR end position error exceeds 0.1 m for single joint motion. The average value of CMOR end position error is less than 0.025 m after compensation, and the absolute error value is reduced by 4 times, improves the dynamic control accuracy of CMOR.

In this paper, an adaptive motion planning method is proposed for the control requirements of the China Fusion Engineering Test Reactor multipurpose overload robot (CMOR). Firstly, a kinematic model of CMOR is established by Denavit-Hartenberg method, and the working process of entering and leaving the vacuum chamber to complete the maintenance operation is analyzed. An iterative tractrix-based motion planning method is proposed by expanding the basic principle of tractrix. Assuming that the CMOR any link has two mutually perpendicular degrees of freedom to achieve spatial motion capability, a more natural and smooth movement of the CMOR is achieved by recursive operations on the tractrix traction. To accelerate the calculation process and avoid each joint rotation angle of CMOR exceeding the set value, a bidirectional iterative tractrix method and a joint limit rotation angle optimization method based on the bidirectional iterative tractrix method are proposed to improve the convergence speed and motion safety. Finally, a CMOR visualization simulation system is built, and the results show that the CMOR adaptive motion planning method can effectively track the target trajectory and avoid joint rotation angle overrun.

Recently, the rapid development of digital twin (DT) technology has been regarded significant in Cyber -physical systems (CPS) promotion. Scholars are focusing on the theoretical architecture and implementing applications, in order to establish a high-fidelity, dynamic, and full-lifecycle DT model and achieve a deep fusion of real and virtual. As a typical complex system with multi-disciplines, multi-physics, and multi-domain characteristics, industrial robot (IR) involves various processes and elements from the two other levels of the system: components and production lines. Their complex relationships lead to a huge challenge to build a comprehensive DT model. Current researchers usually concentrates on single-layer services because of limited construction methodology, which results in enormous isolated models, and leads to low reusable system blocks, finite scalability, and high costs of design, adjustment, upgrade, and maintenance. To address these issues, a standardized methodology and a hierarchical, modular, and generic architecture are proposed to depict comprehensive and variable industrial robot digital twin (IRDT). Firstly, the ontology information model is presented by analyzing variable factors systematically. Then, model-based system engineering (MBSE) based methodology is introduced, including construction process and variants management. After modeling process of three levels (problem domain, solution main, and implementation domain) and four viewpoints (requirement, structure, behavior, and parameter), a generic architecture of IRDT is constructed and a feature-based variants management method is described. Besides, a six-axis IRDTS is implemented to illustrate the mapping of logical architecture and physical system as a multi-level elements and processes representation example. And the steps of numerical evaluations consist of system delay and derivation. Finally, results show the effectiveness and the potential of the proposed theoretical methodology for constructing IRDTS and other industrial applications.

In the field of equipment health management, fault prediction is a valuable and challenging task. Previous fault prediction methods are generally designed for specific mechanical equipment with the specific requirements of data type (e.g., vibration acceleration signals), and they cannot be applied to different mechanical equipment. In this paper, a fault prediction modeling framework based on relationship mining and graph neural network is proposed, which can be widely used in different types of mechanical equipment for fault prediction. First, based on alarm records and maintenance records, the Apriori algorithm is used to analyze the relationship between alarm signals and fault events. Then, the alarm track graph (ATG) for a certain period (e.g., seven days) is constructed based on the key alarm signals. Finally, the graph convolutional network (GCN) is used to predict the fault occurrence for a future certain period (e.g., three days). As can be expected, the modeling approach can effectively mine the correlation between the alarm records and maintenance records, and thus make reliable predictions of future faults. Since the above process utilizes only two common data types, i.e. alarm records and maintenance records, the proposed prediction framework can be easily implemented in a practical background.

基于物联网的全天候实验教学适应了我国高等教育大众化的发展要求,对于培养大学生的动手实践能力,尤其是培养优秀的新型人才具有十分重要的作用.开放实验室智能管理系统及配套的物联网智能硬件终端,能够很好地满足实验室开放过程中教师和学生的实际需求,尤其是能够最大程度解决实验室开放过程中的安全问题,有效促进了实验室及其设施的开放,全天候地保障了实验教学的正常开展.

This article proposes a method to remotely control the robot in the laboratory through the Kinect camera to solve the impact of the Covid-19 epidemic on the laboratory teaching experience which allows users to remotely control robots through their own body movements to understand the principles of robots. It is used to solve the problem of fewer students willing to participate in robot remote education. In this study, the Azure Kinect DK camera was used to collect the motion posture of the upper limbs of the human body. The Kinect camera calculates the frames of human arm joints’ motion. The control system calculates the direction of motion of each joint of the human body based on the quaternion by mapping the heterogeneous human joints with the robot joints. Make the posture of the human arm swing correspond to the posture of the robot’s movement. Thus, the robot in the laboratory can be driven remotely through Azure Kinect DK. By using the method described in this article, students use the camera’s motion capture system to remotely manipulate the robot to grab some simple objects. Through the method described in this research, students can carry out some simple operations on the robots in the laboratory from remote. So it is convenient for students to understand the basic principles of robots and achieve the purpose of better remote experimental teaching. At the same time, students can get practical application of motor servo control, ergonomics, physical simulation engine, digital twin system, etc.

为弥补传统工程培训中培训效率低、危险性高、针对性弱等缺点,提出一种基于数字孪生技术的工程培训模式.在该模式的构建过程中,阐释了在工程培训背景下数字孪生体的内涵,给出了孪生体的具体构建方法.同时,提出了训练平台的设计方案,包括培训软件的总体架构设计、系统通信架构设计、平台资源构建与资源管理方案设计等.最后,以电力系统培训为例给出了该模式的具体实现过程,表明了基于数字孪生工程培训模式应用的可行性.

Abstract In this study, we introduced a machine learning method for estimating human walking speed using plantar pressure and acceleration data. A pressure-derivative method using pretest feature selection was proposed to extract speed-related features from plantar pressure sensors. The maximum, minimum, and standard deviation of acceleration data were also selected as neural network inputs. To improve the generalization ability of the neural network, Bayesian regularization method was adopted. Experiments were conducted under seven different walking speeds to validate the performance of the proposed method. The results show that a strong linear correlation (R = 0.995) exists between the estimated and actual walking speed. The average error of the proposed method is 0.003 ± 0.043 m/s (mean ± root-mean-square error), which is better than previous works. It is suggested that including the speed-related information of both stance and swing phase would give a new insight for achieving a high accuracy of walking speed estimation.

With the development of a new generation of information technology, digital twin generation has become a research trend. However, as a complex system oriented to the product life cycle, there are some challenges for its achievement. As a widely accepted and applied methodology, most industrial fields agreed with that model-based system engineering (MBSE) can improve design quality, development efficiency and avoid risks in the design phase. However, there is few work to discuss the connection between them. Therefore, we explore to apply the model-base system architecture process to implement digital twin system, and adjust some detailed design process for adapting to the characteristics of DT. Finally, taking a human-robot system as an example to illustrate the process of building digital twin system architecture with SysML, moreover, architectural trade-off analysis method (ATAM) demonstrate the applicability of the MBSE-based method in digital twin system designing. This paper presents digital twin system development process based on MBSE based method, and provide a reference for other researchers.

In this study, we introduced a machine learning method for estimating human walking speed using plantar pressure and acceleration data. A pressure-derivative based with pretest feature selection method was proposed to extracted speed-related features from plantar pressure sensors. The maximum, minimum and standard deviation of acceleration data were also selected as neural network inputs. To improve the generalization ability of the neural network, a Bayesian regularization method was adopted. To validate the performance of the proposed method, experiments were conducted under seven different walking speeds. The results show that a strong linear correlation (R = 0.995) exists between the estimated and the actual walking speeds. The average error of the proposed method is 0.003 ± 0.043 m/s (mean ± root mean square error), which is better than previous works. The desirable performance of the proposed method proves that including the speed-related information of both stance and swing phase is beneficial for improving the accuracy of walking speed estimation.

A self-paced treadmill automatically adjusts speed in real-time to match the user’s walking speed, presumably leading to a more nature gait than fixed-speed treadmill. However, previous study has proven that the acceleration applied to the subjects would influence the gait stability. In order to have insights on to which extent will the accelerations affect gait stability, simulation analysis based on conceptual model has been done in the current study. This paper utilized a non-inertial frame based spring-loaded inverted pendulum model to analysis the condition of stability during continuous self-paced treadmill walking. Simulations were done for 100 continuous self-paced treadmill walking at the normal walking speed. And 10ms impulse accelerations of different magnitudes with the range of (−1g, 1g) were applied at different gait events such as toe-off, foot-flat and heel-strike. The simulation results showed that the magnitude of the accelerations had significantly influence on continuous self-paced treadmill walking and directional-dependency was also found. However, no significantly difference was found when applying the impulse acceleration at different gait events.

This study investigates the effects of treadmill control algorithms on spatiotemporal variables when walking on a self-paced (SP) treadmill. Ten healthy subjects walked at their preferred walking speed for 15 min under three different treadmill control modes. Stride time, stride length, and stride speed were measured using an inertial measurement unit. The mean, coefficient of variance, Poincaré descriptors, and gait dynamics were calculated for each parameter. The mean values of stride length and stride speed were significantly increased when the treadmill had a quick response speed to the user’s walking behavior. The long-term variability of stride length and stride speed was significantly affected by the treadmill control algorithms. A reduced strength of long-range correlations of stride time and stride speed was found when walking on the SP treadmill with suppressed treadmill accelerations and small velocity variations. We suggest that the suppression of treadmill acceleration provides more adaptability and less constraint to the user during SP treadmill walking. Although further research is required, the present work provides a basis for interpreting the influence of treadmill control algorithms on human gait.

In this paper, deep learning and model-based method were combined using Bayesian Inference to realize high accuracy and good generalization capability stride-by-stride walking speed estimation by low cost inertial measurement unit (IMU) sensor. Long Short-Term Memory (LSTM) network was applied to train the prediction model because of its ability to consider the temporal correlation of multi-dimensional kinematic parameters during one stride. To improve the performance with unseen subjects, a model-based method was introduced for its relatively good generalization capability. Fusion strategy based on Bayesian Inference was applied to take advantage of the two methods which considered the estimation derived from different methods as abstract sensors. Six healthy young adults performed treadmill walking with shank-mounted IMU and the range of the walking speed was 2.5 km/h to 5 km/h at an increment of 0.5 km/h. Leave-one-subject-out (LOSO) crossvalidation was performed to analyze the generalization capability. For deep learning method, the root mean square error (RMSE) of the model trained by all available subjects was 0.026 m/s and the RMSE of LOSO cross-validation was 0.066 m/s which indicated a low generalization capability. After the fusion strategy was applied, RMSE of the model trained by all available subjects was 0.023 m/s which was slightly improved, while the RMSE of LOSO cross-validation was reduced to 0.036 m/s which indicated that accuracy and the generalization capability was greatly improved. In addition, this accurate estimation can be easily realized online which is essential for locomotion interactive systems (e.g. self-paced treadmill).

自适应跑步机是康复医疗和人体工程学的研究热点,也是虚拟现实运动输入设备的重要组成部分.针对自适应跑步机上人体位置相对地面几乎无变化,很难通过简单的位置差分获得到人体运动速度的问题,提出了一种适应性广、无标记点、非接触式的行走速度估计方法.将Kinect采集到的人体关节点位置数据通过四元数标定、高斯滤波和三次样条插值处理后,用步长修正算法计算出行走时的步态时空参数.基于步态时空参数估计使用者在自适应跑步机上行走时的速度.在固定速度的跑步机上,通过将速度估计值与所设定的跑步机实际速度进行对比,验证了速度估计算法有效性,该速度估计算法可以适用于自适应跑步机的控制算法开发.

The purpose of this study was to assess the influence of gait stability induced by treadmill accelerations during self-paced treadmill walking (SPW). Local dynamic stability of three-dimensional (3D) upper body accelerations and hip angles were quantified. The results demonstrated that SPW was more unstable and had higher risk of falling than fixed-speed treadmill walking (FSW) under the impact of treadmill accelerations. The frequency domain analysis of treadmill speed indicated that intrastride treadmill speed variation was the dominating cause of the instability, and self-paced control strategies which can reduce the intrastride variation may achieve higher gait stability during SPW.