Top Read

    Published in last 1 year |  In last 2 years |  In last 3 years |  All
    Please wait a minute...
    For Selected: Toggle Thumbnails
    Supply Chain Resilience: Research Review and Prospects
    Zhong-zhong JIANG, Jia-run GUO, Wei ZHENG
    Journal of Northeastern University(Natural Science)    2025, 46 (7): 59-70.   DOI: 10.12068/j.issn.1005-3026.2025.20250055
    Abstract3168)   HTML85)    PDF(pc) (2764KB)(4073)       Save

    In recent years, compounded crises such as geopolitical conflicts (e.g., the Russia-Ukraine conflict) and technological containment (e.g., the China-U.S. trade friction) have continuously exerted a profound impact worldwide, revealing the vulnerabilities of global supply chains. Enhancing the supply chain resilience has become a critical strategy to ensure the sustainable development of countries around the world, especially China, and it serves as a vital foundation for making China strong in manufacturing. On this basis, existing research on supply chain resilience was comprehensively reviewed, with particular focus on its origins, conceptual definitions, and driving factors. The evolution of the research was systematically analyzed, and prospective research directions were explored from four dimensions: collaborative optimization, resource allocation, dynamic response, and risk management. The findings aim to provide theoretical support and decision reference for enhancing supply chain resilience both globally and within China.

    Table and Figures | Reference | Related Articles | Metrics | Comments0
    Development and Prospects for Software‑Defined Intelligent Control Systems
    Tian-you CHAI, Rui ZHENG, Yao JIA, Xin-yu HUANG, Yan-jie SONG
    Journal of Northeastern University(Natural Science)    2025, 46 (7): 1-10.   DOI: 10.12068/j.issn.1005-3026.2025.20250079
    Abstract1642)   HTML85)    PDF(pc) (4009KB)(990)       Save

    The current state of research on software-defined control systems was reviewed, and the role and development of control systems throughout the industrial revolutions were analyzed. The intelligent development direction for software-defined control systems was proposed. The case study of a software-defined end-edge-cloud collaborative PID(proportional-integral-derivative) tuning intelligence system was presented, which demonstrates that the tight conjoining and coordination between industrial artificial intelligence, industrial Internet, and other new-generation information technologies with software‐defined control systems has opened up a new way for the development of software-defined intelligent control systems. Finally, the principal research directions for software-defined intelligent control systems were pointed out by considering the challenges faced by software-defined control systems and those specific to their intelligent transformation.

    Table and Figures | Reference | Related Articles | Metrics | Comments0
    Research Progress on Development and Application of Digital Blast Furnace Ironmaking Technology
    Man-sheng CHU, Guo-dong WANG, Jue TANG, Quan SHI
    Journal of Northeastern University(Natural Science)    2025, 46 (7): 113-130.   DOI: 10.12068/j.issn.1005-3026.2025.20250070
    Abstract1351)   HTML16)    PDF(pc) (4726KB)(4036)       Save

    With the advancement of the digital information era, the digital transformation of blast furnaces has begun. Steel enterprises have applied intelligent closed-loop control, digital twins, and AI-based predictive models to develop intelligent systems for smart blast furnace operation, blast furnace condition assessment, and quality optimization. Research on digital blast furnaces primarily focuses on variable prediction, state diagnosis, and blast furnace condition optimization, with these domains evolving from traditional approaches toward complex optimization modeling, multidimensional comprehensive evaluation, and multi-objective collaborative optimization, respectively. However, current predictive models require enhanced online self-updating and integration of data and mechanisms; evaluation systems need to emphasize multidimensional and fine-grained diagnostics, and blast furnace condition optimization has to overcome single-indicator limitations by focusing on low-risk, low-cost, and multi-objective coupled strategies. According to the actual needs of the blast furnace site, a physical system of blast furnace information was developed, where data, mechanisms, and experience were reasonably matched and called upon to form an integrated technology encompassing data governance, rule mining, intelligent prediction, comprehensive evaluation, multi-objective optimization, and decision feedback, which was identified as one of the key directions for future development of digital blast furnace ironmaking.

    Table and Figures | Reference | Related Articles | Metrics | Comments0
    Progress and Application of Intelligent Manufacturing Technology
    Ya-dong GONG, Jia-hao GAO, Li-ya JIN, Heng ZHAO
    Journal of Northeastern University(Natural Science)    2025, 46 (7): 84-93.   DOI: 10.12068/j.issn.1005-3026.2025.20250053
    Abstract1351)   HTML44)    PDF(pc) (4156KB)(5112)       Save

    Based on an analysis of the extensive and profound impact of the development of artificial intelligence on manufacturing technology, it is expounded that intelligent manufacturing is the main direction for building a strong manufacturing country, and its important roles are discussed in the construction of a new industrialization system. Through an introduction to intelligent manufacturing empowerment technology, it is claimed that intelligent manufacturing technology by the deep integration of intelligent technology and advanced manufacturing technology will drive the progress of the new round of industrial revolution and industrial transfor-mation. Subsequently, typical research results and application examples of intelligent manufacturing are used to clarify the essential differences and engineering applications between intelligent manufacturing and traditional automation. Finally, it is pointed out that the advanced production mode of the deep integration of new generation information technology and advanced manufacturing technology demonstrates a new form of future manufacturing industry.

    Table and Figures | Reference | Related Articles | Metrics | Comments0
    Review of Multi-type Energy Routers Research
    Qiu-ye SUN, Rong-da XING, Qian-xiang SHEN, Zhen-ao SUN
    Journal of Northeastern University(Natural Science)    2025, 46 (7): 11-21.   DOI: 10.12068/j.issn.1005-3026.2025.20250063
    Abstract1317)   HTML20)    PDF(pc) (1261KB)(1094)       Save

    Energy routers (ERs) are one of the core components of the energy Internet for achieving multi-port energy conversion and active energy flow control. This paper classified ERs into three categories: electrical ERs, information ERs, and multi-energy ERs. Based on the differences between these categories, the research on ERs is divided into four aspects: electrical conversion, focusing on topology and control of multi-port electrical conversion; energy routing control, primarily concerned with the regulation of power flow between ports of ERs; information processing and optimal control, emphasizing the acquisition and transmission of information and optimizing energy flow; and multi-energy coordination, with multi-energy comprehensive utilization as the main goal. Based on these four research aspects, this paper explored topology, control, communication, and multi-energy optimization of ERs, as well as the interrelationships between different aspects.

    Table and Figures | Reference | Related Articles | Metrics | Comments0
    Intelligent Identification Method of Industrial Mixed Gases Based on ConvGRU Fusion Attention Mechanism
    Fan-li MENG, Shu-chang LI, Hao WANG, Zhen-yu YUAN
    Journal of Northeastern University(Natural Science)    2025, 46 (7): 37-48.   DOI: 10.12068/j.issn.1005-3026.2025.20240164
    Abstract1172)   HTML22)    PDF(pc) (4835KB)(563)       Save

    To address the issue of high data dependency and insufficient accuracy in mixed gas identification for traditional semiconductor gas sensors, a ConvGRUAttention network model that integrates gated recurrent units (GRU), convolutional layers, and attention mechanism is proposed. Empirical wavelet transform (EWT) is employed to convert raw signals into the time-frequency domain and perform multi-scale decomposition, which suppresses noise, reduces data dependency, and enhances the model’s robustness. The model extracts local dynamic features through convolutional layers, captures long-term dependencies using GRU, and optimizes feature weights across multi-scale signals via the attention mechanism, thereby improving feature extraction and generalization capabilities. Experimental results demonstrate 100% accuracy in qualitative identification and a root mean square error (RMSE) of 3.3×10⁻⁶ in quantitative detection. Compared with the traditional methods, the detection accuracy for mixed gases is significantly improved.

    Table and Figures | Reference | Related Articles | Metrics | Comments0
    Research on Localization of Industrial Intelligent Inspection Robots in Cable Tunnel Environment
    Yu-tao WANG, Jun-wei AN, Chang-sheng QIN, Wei-fan GUO
    Journal of Northeastern University(Natural Science)    2025, 46 (7): 49-58.   DOI: 10.12068/j.issn.1005-3026.2025.20240212
    Abstract1096)   HTML17)    PDF(pc) (5943KB)(163)       Save

    The cable tunnel is closed and narrow, with repetitively laid cable racks and similar scene textures, which is a degraded scenario. To address this environment, a visual-inertial SLAM (simultaneous localization and mapping) algorithm based on point-line feature fusion is proposed. The algorithm improves the high-dimensional line features through length suppression and short line fitting to make it more effective in describing the structural features of tunnel scene. In addition, for the problem of loop closure detection failure due to feature similarity in cable tunnels, ArUco markers with efficient recognition and accurate pose estimation are introduced to limit the loop closure area, and the optimal loop closure frames are selected using the minimized pose transformation to improve detection accuracy and localization precision. Finally, dataset collection and experimental validation were conducted in actual cable tunnels. The results show that the absolute trajectory accuracy of the algorithm is improved by 69.73% on average relative to VINSMono(visual intertial system-Mono), which meets the application requirements of cable tunnel inspection.

    Table and Figures | Reference | Related Articles | Metrics | Comments0
    Digital Twin Fault Diagnosis Method of Power Transformer Based on Industrial Intelligence
    Jian FENG, Bo-wen ZHANG, Ning ZHAO, Hui-jie JIANG
    Journal of Northeastern University(Natural Science)    2025, 46 (7): 22-29.   DOI: 10.12068/j.issn.1005-3026.2025.20240218
    Abstract1092)   HTML18)    PDF(pc) (4393KB)(336)       Save

    As a key development direction integrating new-generation information technology with advanced manufacturing techniques, industrial intelligence leverages intelligent, digital, and automated methods to significantly enhance industrial production efficiency and optimize the prediction and maintenance management of industrial equipment. This paper focuses on the intelligentization of industrial equipment, with the goal of ensuring the efficient and stable operation of power transformers within power systems. A digital twin model for transformer inter-turn short circuit faults is constructed based on electromagnetic field equations and equivalent circuit models. The model analyzes the symmetry of the transformer in both normal and fault conditions from an electromagnetic field perspective, thereby integrating digital twin technology with fault diagnosis. Furthermore, through in-depth analysis of the virtual model of the transformer, the location of faults is accurately identified, ensuring the safe operation of the transformer, improving its reliability and efficiency, and advancing the intelligentization and modernization of the entire power system.

    Table and Figures | Reference | Related Articles | Metrics | Comments0
    Intelligent Microseismic Monitoring and Early Warning for Rock Burst During TBM Excavation of Deep Metal Mines
    Bing-rui CHEN, Xu WANG, Gui-peng JIANG, Fei HE, Jia-lin HAN, Jian-jun HAO
    Journal of Northeastern University(Natural Science)    2025, 46 (7): 148-162.   DOI: 10.12068/j.issn.1005-3026.2025.20250085
    Abstract1029)   HTML15)    PDF(pc) (10803KB)(1701)       Save

    In response to the problem of insufficient automation and intelligence in the microseismic monitoring and early warning for rock bursts during tunnelling boring machine (TBM) excavation of deep metal mines, research on multi-dimensional parameter recognition of drilling holes based on deep machine vision DPED-AT method was conducted; automatic disassembly and assembly device for microseismic sensors was developed, and the decision-making system was designed, achieving automatic disassembly and assembly of microseismic sensors during TBM excavation. Microseismic intelligent frequency conversion acquisition technology was developed, realizing continuous and high-fidelity acquisition of rock rupture information during the rock burst incubation process. An improved neural network algorithm for identifying and picking up rupture signals was proposed, as well as a three-dimensional characterization algorithm for the probability field of microseismic sources induced by rock bursts incubation. Intelligent interpretation and refined early warning of rock burst incubation information during TBM excavation were preliminarily realized, and an intelligent monitoring and early warning technology system for rock burst that integrated intelligent drilling hole recognition, automatic sensor disassembly and assembly, and intelligent signal acquisition and interpretation was ultimately established. The application in Ruihai Gold Mine shows that it has achieved automatic microseismic monitoring, interpretation, and early warning of rock burst, providing strong support for less manned and unmanned TBM excavation in deep metal mines.

    Table and Figures | Reference | Related Articles | Metrics | Comments0
    Charging and Discharging Scheduling for Electric Vehicles Based on Improved Multi-objective Chaotic Particle Swarm Optimization
    Zhi-ao CAO, Chen-shuo MA
    Journal of Northeastern University(Natural Science)    2025, 46 (9): 1-8.   DOI: 10.12068/j.issn.1005-3026.2025.20240014
    Abstract917)   HTML67)    PDF(pc) (1428KB)(401)       Save

    To address the issue of charging and discharging scheduling for EVs(electric vehicles), an orderly charging and discharging algorithm that considered users’ comprehensive satisfaction was proposed. Firstly, a large-scale orderly charging and discharging model for EVs was constructed, and users’ comprehensive satisfaction was quantified. Secondly, an improved multi-objective role partitioning chaotic particle swarm optimization(IMRPC-PSO) algorithm was proposed to solve the problems of insufficient diversity and being trapped in local optimal in traditional methods. According to the performance of particles, the roles of elite particles, general particles, and learning particles were assigned, which respectively implement diversity strategies of maintaining search, developing search, and learning search. Each particle searched the optimization space according to its assigned role. To avoid falling into local optimal, a chaotic sequence perturbation was added after the initialization of each iteration. Finally, the performance of the proposed algorithm was compared with that of the other five multi-objective optimization algorithms through case simulation. The results show that IMRPC-PSO is superior to other algorithms in solving the problem of orderly charging and discharging of EVs, verifying the effectiveness and feasibility of the proposed algorithm.

    Table and Figures | Reference | Related Articles | Metrics | Comments0
    Analytical Modeling and Performance Evaluation of Multi-stage Assembly Lines with Line-Side Buffers
    Peng-hao CUI, Qi-man ZHANG, Zhong-zhong JIANG, Guo-jun SHENG
    Journal of Northeastern University(Natural Science)    2025, 46 (7): 71-83.   DOI: 10.12068/j.issn.1005-3026.2025.20240193
    Abstract909)   HTML5)    PDF(pc) (2384KB)(555)       Save

    The output performance of assembly lines is not only affected by machine unreliability and limited buffer capacity but also constrained by line-side buffers. The analytical modeling and performance evaluation of multi-stage assembly lines with line-side buffers were investigated. Firstly, for the single-stage assembly lines, the steady-state probability distribution of system states was derived based on Markov chains. Secondly, for the two-stage assembly lines, each single-stage subsystem was modeled as a machine with one operational state and one failure state. A performance evaluation model was then established using Markov chains, and closed-form expressions for performance indicators were obtained. Thirdly, for the multi-stage assembly lines, an aggregation method was proposed to approximate the performance indicators. Furthermore, the accuracy of the performance evaluation method was validated through numerical experiments. Finally, utilizing the proposed method, numerical experiments were conducted to examine system properties, such as reversibility and monotonicity in the multi-stage assembly lines.

    Table and Figures | Reference | Related Articles | Metrics | Comments0
    Efficient Semi-supervised Medical Image Lesion Segmentation Method Based on MedSAM
    Xi-bin JIA, Xun-jie YIN, Chao FAN, Zheng-han YANG
    Journal of Northeastern University(Natural Science)    2026, 47 (1): 1-10.   DOI: 10.12068/j.issn.1005-3026.2026.20259022
    Abstract906)   HTML57)    PDF(pc) (1940KB)(754)       Save

    In semi-supervised lesion segmentation, the performance of the teacher network is poor, making it difficult for it to guide the student network to perform effective segmentation. To address this issue, an efficient semi-supervised medical image lesion segmentation method was proposed, employing the medical segment anything model (MedSAM), which exhibited superior feature extraction capabilities, as the teacher network. A lightweight student network based on Mamba was constructed, and its segmentation performance was enhanced through knowledge distillation. To address the semantic mismatch caused by feature alignment across heterogeneous networks, a perturbation-consistent cross-architecture knowledge distillation method was introduced. This approach mapped teacher features to the student feature space and aligned perturbation responses, thereby improving the student network’s feature representation ability and improving segmentation performance. Additionally, to tackle the challenges of diverse lesion morphologies and low foreground-background contrast, leading to poor segmentation consistency, a distribution-based self-supervised loss was proposed for optimization. Experiments on multiple types of medical image lesion segmentation datasets demonstrate that the proposed method in this paper outperforms existing methods in segmentation performance. Meanwhile, the student network has only 1.34 M parameters, which significantly improves the model efficiency.

    Table and Figures | Reference | Related Articles | Metrics | Comments0
    Defect Detection on PCB Based on Improved YOLOv8
    Zhen-zhen LYU, Li-jin FANG, Qian-kun ZHAO, Ying-cai WAN
    Journal of Northeastern University(Natural Science)    2025, 46 (10): 1-9.   DOI: 10.12068/j.issn.1005-3026.2025.20240038
    Abstract901)   HTML37)    PDF(pc) (10197KB)(152)       Save

    Due to the high integration, complex circuits, and increasing parameters of printed circuit boards (PCBs), defects in PCBs directly affect production efficiency, making computer vision-based defect detection crucial for PCB manufacturing. A self-attention-based PCB defect detection algorithm was proposed based on the YOLO object detection algorithm. First, a polarized self-attention (PSA) mechanism was introduced in the feature extraction stage to separately extract spatial and semantic features of PCBs, which were combined with input raw features to enhance the network’s feature representation capability. Then, a small-object detection head was added in the decoding stage, which fully utilized low-resolution features from the YOLO network Backbone module to enable the network to focus on local details of PCBs and improve defect positioning accuracy. Experiments show that the proposed method achieves 95.5% accuracy on the PCB dataset, 4% higher than the original YOLOv8 method, with the mAP0.5∶0.95 metric increased by 2.8%.

    Table and Figures | Reference | Related Articles | Metrics | Comments0
    Weibull Distribution Parameter Estimation Method Based on Statistical Minimum Diversity Principle
    Li-yang XIE, Wen-hui ZHU, Ning-xiang WU, Xiao-yu YANG
    Journal of Northeastern University(Natural Science)    2025, 46 (7): 108-112.   DOI: 10.12068/j.issn.1005-3026.2025.20240194
    Abstract812)   HTML8)    PDF(pc) (1860KB)(604)       Save

    For the Weibull distribution parameter estimation, a pseudo-estimator of scale parameters is constructed, and the estimated parameter values can be obtained by finding the extreme point of relevant variables based on the principle that the right location parameter and shape parameter minimize the diversity of the scale parameter estimates associated with individual sample values. Essentially, parameter estimation extracts(overall)information based on specific patterns reflected by a set of data with uncertainty (random variable samples). However, the pattern is statistical in nature rather than deterministic. In terms of the occurrence of extreme points in the related functions, the exact value of the estimater does not necessarily occur at the extreme point in a deterministic extreme point. It is shown that there is typically a deviation between the point where the exact parameter is located and the theoretical extreme point, and the accuracy and robustness of the parameter estimation method can be greatly improved by introducing an offset value in the minimum value criterion (modifying “the first derivative being equal to zero” to “the first derivative being equal to a value greater than zero”). A large number of parameter estimation cases show that the range of the estimated value of the Weibull location parameter (true value is 1 000) is narrowed from 0~1 500 to 500~1 550 by taking an offset value of 0.1.

    Table and Figures | Reference | Related Articles | Metrics | Comments0
    Risk-Oriented Crowd Navigation Strategy Based on Deep Reinforcement Learning
    Yang JIANG, Tian-xiang ZHAO, Ruo-huai SUN, Lei WANG
    Journal of Northeastern University(Natural Science)    2025, 46 (12): 1-8.   DOI: 10.12068/j.issn.1005-3026.2025.12.20240094
    Abstract779)   HTML25)    PDF(pc) (2715KB)(431)       Save

    To improve robot freezing and suboptimal performance of traditional navigation methods in the presence of dynamic obstacles, a navigation method based on deep reinforcement learning was proposed. The core of this method lies in its risk perception module and path selection module. The risk perception module calculated the collision probability between the robot and nearby dynamic obstacles in real time, allowing the robot to prioritize avoiding more hazardous obstacles. Concurrently, the path selection module evaluated the “passing ability” of the robot in surrounding areas in real time, guiding the robot to choose safer paths. In comparison experiments with a deep reinforcement learning method that lacks these modules, the proposed method achieved the highest navigation success rate in all simulation test environments, with an improvement rate of up to 11%.

    Table and Figures | Reference | Related Articles | Metrics | Comments0
    Learning-Based NSGA-Ⅱ for Multi-objective Portfolio Optimization Problems
    Zhu ZHU, Hang-yu LOU
    Journal of Northeastern University(Natural Science)    2025, 46 (9): 17-24.   DOI: 10.12068/j.issn.1005-3026.2025.20249030
    Abstract764)   HTML10)    PDF(pc) (1542KB)(1204)       Save

    To address the issues of insufficient diversity and poor constraint-handling capability in the non-dominated sorting genetic algorithm Ⅱ (NSGA-Ⅱ) when solving portfolio optimization problems, a learning-based improved NSGA-Ⅱ algorithm (INSGA-Ⅱ) incorporating clustering and an adaptive feasibility repair strategy for multi-objective portfolio optimization was proposed. In the proposed algorithm, clustering learning was employed to enhance population diversity, while adaptive repair ensured that newly generated solutions were feasible, thereby improving the algorithm's diversity and convergence speed. Additionally, the populations after crossover and mutation were preserved separately and merged with the parent population to increase the selection pressure and quality of offspring generation. Experimental results demonstrate that the proposed algorithm exhibits superior search performance and stability, effectively solving multi-objective portfolio optimization problems.

    Table and Figures | Reference | Related Articles | Metrics | Comments0
    An Improved Small Object Detection Model Based on YOLOv8 for UAV Vision
    Ji-hong LIU, Rui-rui SHI
    Journal of Northeastern University(Natural Science)    2025, 46 (12): 29-37.   DOI: 10.12068/j.issn.1005-3026.2025.20240116
    Abstract761)   HTML24)    PDF(pc) (10174KB)(1279)       Save

    In view of easy false detection and missed detection of small objects in unmanned aerial vehicle (UAV) aerial images, as well as the requirements for real-time performance and lightweight design in UAV detection tasks, an improved lightweight and efficient model based on YOLOv8 was proposed. Firstly, the Neck part of YOLOv8 was simplified into a feature pyramid network, enabling the model to effectively utilize the detailed information extracted by shallow networks. Meanwhile, a feature fusion module was added to provide more favorable features for small object detection to the Head layer. Secondly, an efficient local attention (ELA) mechanism was integrated into the Backbone part to achieve accurate localization of target regions. Experimental results show that compared with YOLOv8s, the parameters and model size of the improved model are reduced by 50%, while the mAP0.5 and detection speed are improved by 4%. This improved model provides a new idea for the deployment of UAV detection.

    Table and Figures | Reference | Related Articles | Metrics | Comments0
    Optimal Control of Hydrogen Production by Renewable Energy Source Considering State of Charge of Energy Storage
    Zhi-liang WANG, Liang-liang GUO, Xin-yu LI, Xin-rui LIU
    Journal of Northeastern University(Natural Science)    2025, 46 (7): 30-36.   DOI: 10.12068/j.issn.1005-3026.2025.20240184
    Abstract750)   HTML11)    PDF(pc) (1978KB)(417)       Save

    An optimal control strategy considering the state of charge (SoC) of energy storage is proposed for an isolated DC microgrid for hydrogen production system composed of renewable energy, electrolytic cell, and energy storage equipment. Firstly, the characteristics of hydrogen production efficiency of alkaline electrolyzers are analyzed, and an optimal control method for adaptive adjustment of hydrogen production efficiency with bus voltage change is proposed. By coordinating with the energy storage system, the hydrogen production efficiency is kept within a high range. When the SoC of energy storage violates the upper and lower limits, a communication-independent SoC active recovery control strategy is designed to ensure the safe operation of the energy storage system. Secondly, a power coordinated control strategy considering extreme conditions is designed to ensure the stable operation of the DC microgrid through flexible switching between various operating modes. Finally, the effectiveness of the proposed control strategy is verified by MATLAB/Simulink simulation platform.

    Table and Figures | Reference | Related Articles | Metrics | Comments0
    Design of Nonlinear Observer for Distributed Drive Electric Vehicle with Actuator Faults
    Hong-wei WANG, Xin-yu JI, Jin-shuo SONG
    Journal of Northeastern University(Natural Science)    2026, 47 (1): 60-66.   DOI: 10.12068/j.issn.1005-3026.2026.20240152
    Abstract726)   HTML11)    PDF(pc) (1352KB)(145)       Save

    A robust fault diagnosis algorithm based on an observer was designed for the nonlinear dynamic system of distributed drive electric vehicles with actuator faults. Firstly, by considering the actuator fault and system disturbance, a mathematical model of 7 degrees of freedom vehicle system was established. Then, a fault diagnosis strategy based on a nonlinear observer was designed by using linear matrix inequality and Lyapunov function to realize multiple estimations of system state, disturbance information, and fault factor, and the relevant theoretical proof was given. Finally, by referring to the actual driving conditions of the vehicle, the co-simulation was carried out by MATLAB/Simulink and CarSim. The results show that the designed observer can accurately diagnose and estimate the actuator faults, rapidly track the actual operating state of the system, and improve the anti-interference ability of the system.

    Table and Figures | Reference | Related Articles | Metrics | Comments0
    Two-Stage SiamCAR Tracking Algorithm Combining Motion Information and Dual-attention Mechanism
    Ying WEI, Jia-peng ZHANG, Jia-qi CUI, Tong HUANG
    Journal of Northeastern University(Natural Science)    2025, 46 (9): 9-16.   DOI: 10.12068/j.issn.1005-3026.2025.20240018
    Abstract710)   HTML14)    PDF(pc) (2509KB)(161)       Save

    In single-object tracking, the accuracy of the tracking bounding box is often compromised by factors such as deformation, motion blur, occlusion, and background interference. In particular, background interference frequently leads to tracking hopping and drift. To mitigate these issues, a two-stage tracking algorithm that integrated motion information with a dual-attention mechanism was proposed. In the first stage, a SiamCAR tracker with a dual-attention mechanism was employed to coarsely locate the target in the current frame. In the second stage, a refinement module of the bounding box was constructed using pixel-level similarity computations to learn the subtle features of the target under low-latency conditions, thereby enhancing the tracking accuracy. Finally, the tracking box obtained based on appearance features was fused with the target’s motion trajectory information to mitigate tracking drift and hopping. Experimental results on the OTB100 dataset indicate that the success rate and accuracy of the tracking box have improved by 4.6% and 2.8%, respectively, compared to the original. The success rate in the presence of background interference has reached 69.6%.

    Table and Figures | Reference | Related Articles | Metrics | Comments0