WANG Aoyu. Lost in Definition: How Confusion over Agentic AI Risks Undermining U.S. Governance Frameworks[J]. Technology of IoT&AI, 2026, 58(4): 1-5.
More people in the tech world now call Artificial Intelligence (AI) systems that can plan, act on their own, handle multi-step tasks, and need little human oversight “agentic AI”.Right now, this idea is mostly defined by the industry itself. This report argues that the government should take a more active role in defining it, instead of leaving it entirely to companies.It also warns that unclear definitions can create real problems, making it harder to buy, test, and regulate AI systems. To fix this, the report suggests a new way to look at AI. Instead of only asking “what can it do,” we should also ask how it changes decision-making, power, and responsibility inside organizations. Finally, the report says that if we focus on how AI works in real situations—not just marketing claims—the U.S. can better manage human-AI cooperation, especially in national security, and improve how these systems are bought, evaluated, and governed.
CAI Jincheng. Application of Internet of Things and Deep Learning in Smart Agriculture[J]. Technology of IoT&AI, 2026, 58(4): 6-9.
Smart agriculture, as an important development direction of agricultural modernization, provides strong support for achieving sustainable agricultural development. Firstly, the concepts of the Internet of Things and deep learning are introduced. Secondly, from three core areas: intelligent control of agricultural production environment, full-cycle cultivation management of crops, and allocation of agricultural resources and production equipment, the specific application forms, practical processes, and field application effects of the integration of Internet of Things and deep learning are analyzed, covering specific contents such as dynamic monitoring of soil and meteorological conditions, prediction of environmental risks, identification of crop growth conditions, green prevention and control of pests and diseases, precise application of water and fertilizer, and collaborative operation of agricultural machinery. Finally, the actual value of the technology integration is summarized, existing application shortcomings are sorted out, and subsequent optimization directions are proposed, with the aim of reducing the excessive reliance of traditional agriculture on human experience, achieving data-driven and precise control throughout the agricultural production process, effectively improving the efficiency of agricultural resource utilization and the quality of crop output, promoting the development of agricultural production towards a green and sustainable direction, and providing technical support for the construction of agricultural modernization and the advancement of rural revitalization.
XU Wei. Fault Prediction Method for Thermal Power Plant Equipment Based on Internet of Things[J]. Technology of IoT&AI, 2026, 58(4): 10-14.
Thermal power plant equipment is subjected to high load, high temperature and high pressure environments for a long time, and its operating status is affected by the coupling effect of multiple physical quantities. The evolution of faults is hidden and sudden, so conducting data-driven prediction based on the Internet of Things has important engineering value. Based on this, the fault characteristics of typical thermal power plant equipment are analyzed, and the multi-source data collection and processing strategies and state feature extraction methods in the Internet of Things environment are studied. An intelligent prediction model construction mechanism for complex operating conditions is proposed, and method validation is carried out in actual scenarios. The aim is to build a predictive system for early warning of key equipment in thermal power plants, improve operational reliability and fault diagnosis efficiency. Experimental results show that this method can significantly improve recognition accuracy, early warning amount, and overall operation and maintenance level.
ZHU Jiang. Design of Smart Bedside System Integrating Multimodal Perception and Dynamic Priority Scheduling for Smart Ward[J]. Technology of IoT&AI, 2026, 58(4): 15-21.
In the ward nursing scenarios, the changes in patients’ conditions are sudden and uncertain, making a single calling method difficult to meet the response requirements under different disease states, resulting in delayed emergency call responses, disorder in the scheduling of nursing resources, and mismatch between task allocation and the actual needs of patients. To address this, a smart bedside system integrating multimodal perception and dynamic priority scheduling is designed. At the multimodal perception level, four calling methods including touch screen, voice, automatic warning, and portable remote control are integrated, combined with body sign sensing, electronic medical records, and nurse annotations for real-time identification of patients’ conditions, achieving dynamic configuration of calling methods based on disease severity levels; at the dynamic priority scheduling level, multi-modal fusion positioning technology of Ultra Wide Band (UWB) + Long Range Radio (LoRa) + infrared is introduced, and a scheduling algorithm considering factors such as the urgency of the call, the position and load of nurses, and professional matching degree is constructed to complete intelligent assignment of nursing tasks and resource optimization; dual-mode redundancy and local fault-tolerant mechanisms are designed to ensure the continuous operation capability of the system in complex network environments. Experimental results show that the proposed system controls the average response time of emergency and ordinary calls within 1.5 s, effectively achieving efficient matching of patients’ conditions and nursing resources.
LIU Xi. Research on Internet of Things Based Condition Monitoring Technology for Electromechanical Equipment[J]. Technology of IoT&AI, 2026, 58(4): 22-25.
Addressing the issues of single perception dimension, poor data real-time performance, and low credibility in traditional electromechanical device status information collection, a multi-level status information collection technology system based on the Internet of Things is proposed. This system integrates key technologies such as edge computing preprocessing, Principal Component Analysis (PCA) for feature dimensionality reduction, Message Queuing Telemetry Transport (MQTT) for secure transmission, and blockchain for certificate storage, constructing a complete technical chain from signal acquisition to cloud analysis. Experimental results show that Scheme B achieves a data collection accuracy rate of 97.6%, an improvement of approximately 4.8% compared to the single-sensor scheme; the average response delay is 4.2 s, a reduction of about 51.7%; the feature extraction accuracy rate is 96.3%; in scenarios with strong electromagnetic interference and network disconnection, the data recovery success rate and data tampering detection rate both reach 100%, providing effective technical support for intelligent operation and maintenance of industrial electromechanical devices.
SUN Yongqiang, GUO Yan. Research on Application of Artificial Intelligence of Things Technology in the Electrical Design of Smart Distribution Systems#br#[J]. Technology of IoT&AI, 2026, 58(4): 26-30.
Smart distribution systems require high precision in electrical parameter sensing and dynamic response. Traditional designs are short of the capacity to integrate multi - source loads and forecast faults. This paper analyzes system requirements, identifies structural bottlenecks in parameter acquisition, fault identification, and network optimization, and proposes an Artificial Intelligence of Things (AIoT) application pathway: multi-source load perception reconstruction, artificial intelligence based fault diagnosis, intelligent topology optimization, and platform-coordinated control. This leads to a comprehensive, prompt, and exact smart distribution system, offering innovative support for electrical design.
HAN Chong. Design and Implementation of Intelligent Transportation Internet of Things Architecture Based on 5G[J]. Technology of IoT&AI, 2026, 58(4): 31-34.
To address the issues of scattered access of multi-source devices, insufficient data transmission coordination, and low business linkage efficiency in intelligent transportation, a 3 layer Internet of Things architecture based on 5G is proposed. This architecture consists of the perception access layer, 5G bearer coordination layer, and platform application layer, which work together to achieve data collection, efficient transmission, and intelligent decision-making. The results show that this architecture can improve access integrity, transmission real-time performance, and platform coordination capabilities.
WANG Yao. Research on Operation Control of Microgrid Based on Multi-agent Collaboration[J]. Technology of IoT&AI, 2026, 58(4): 35-38.
To address the rapid dynamics and complex coupling requirements exhibited by microgrid operation under the background of high penetration of distributed power sources, taking multi-scenario hardware-in-the-loop testing as an example, the impact of agent collaborative control on power distribution accuracy, frequency and voltage stability, and dynamic recovery capability is analyzed. The results show that this collaborative control system has significant advantages in terms of voltage deviation suppression, consistency convergence speed, and disturbance robustness, and can effectively improve the autonomous operation quality of microgrids dominated by distributed power sources.
TENG Xiaohui, LIU Faming. Design of Protective Storage Buffer Structure for Shipborne Voyage Date Recorder Based on Cloud Model[J]. Technology of IoT&AI, 2026, 58(4): 39-42.
Aiming at the complex impact environment faced by the protective storage body of the shipborne voyage data recorder under extreme sea conditions, a multi-layer hierarchical buffer structure design method based on the cloud model is proposed. By quantifying the uncertainty of material parameters and loads through the cloud model, the impact response is transformed into predictable statistical characteristics, thereby guiding the integrated configuration optimization of gradient aluminum foam and intelligent damping materials. Experimental results show that this method effectively reduces the peak overload, enhances the structural robustness and energy dissipation efficiency.
LIN Bin. Intelligent Transformation and Collaborative Optimization of Industrial Internet for Power Plant DCS[J]. Technology of IoT&AI, 2026, 58(4): 43-47.
In response to the issues of fragmented control architecture, weak cross-system collaboration capabilities, and insufficient dynamic regulation accuracy exposed by traditional power plant Distributed Control Systems (DCS) in the context of intelligent and clean transformation, an intelligent transformation and collaborative optimization method for industrial internet oriented to power plant DCS is proposed. This method intelligently upgrades key subsystems such as the Modulating Control System (MCS), Sequence Control System (SCS), and Furnace Safeguard Supervisory System (FSSS), thereby constructing a multi-system collaborative optimization model and introducing a particle swarm optimization algorithm for dynamic and balanced load distribution among generating units. The application results show that after the transformation, the coal consumption for power generation of the units has been significantly reduced, the fluctuation of main steam pressure has been effectively suppressed, and both the balance degree of load distribution and the success rate of control command execution have been significantly improved. This work provides a feasible technical path for power plants to explore safe, efficient, and low-carbon operation modes.
WU Dong. Coal Mine Gas Concentration Prediction and Alarm System Based on Machine Learning[J]. Technology of IoT&AI, 2026, 58(4): 48-51.
In response to the problem of strong fluctuation in coal mine gas concentration and the lack of trend prediction capability in traditional threshold alarm methods, a gas concentration prediction and alarm system based on machine learning was constructed. The system integrates functions of multi-source data access, preprocessing, concentration prediction, and alarm linkage. The prediction and warning effects of the system were verified through experiments. The results show that the proposed system can improve the accuracy of gas concentration prediction and the timeliness of alarm, providing technical support for coal mine safety monitoring and risk warning.
WANG Jian, QIAO Shouming. Privacy Protection Data Fusion Algorithm for Industrial Internet of Things[J]. Technology of IoT&AI, 2026, 58(4): 52-55.
To achieve a data processing system that integrates security constraints and fusion capabilities in the industrial Internet of Things environment, a privacy protection data fusion framework is constructed by adopting methods such as hierarchical feature compression, privacy perturbation, and encrypted collaborative update. The quantitative relationship between privacy exposure, fusion error, and system overhead under different deployment densities and feature structures is studied. The multi-stage processing link can maintain stable fusion accuracy under controlled perturbation budgets and enable the collaborative mechanism to maintain scalable operational characteristics in complex topologies.
XIA Zhengxia, WANG Jun, CHEN Yuanling, CAO Dianjie, CHEN Yueming. Research on Automated Orthodontic Technology Based on Interdental Relationship Features[J]. Technology of IoT&AI, 2026, 58(4): 56-62.
Traditional automatic tooth arrangement methods primarily rely on geometric rules and empirical templates. These methods have limited data applicability, complex formulation processes, and are highly dependent on the clinical experience of orthodontists. To address the issues of strong clinical experience dependence and insufficient adaptability in traditional methods, this paper proposes an end-to-end automated tooth arrangement model, PointNet-s+ Bidirectional Long Short-Term Memory (Bi-LSTM). This model utilizes PointNet as the backbone network to extract local and global spatial features of the dental point cloud, and extracts inter-tooth relationship features through the Bi-LSTM structure, significantly improving the rationality and accuracy of the dental arch curve. Additionally, during the data augmentation stage, by analyzing the absence of homonymous teeth, a tooth filling augmentation strategy is proposed. Combined with the Bi-LSTM’s ability to model inter-tooth relationships, this enhances the model’s adaptability to molar mesial movement treatment strategies in scenarios with missing teeth. On the validation set of the ISICDM-ATRC orthodontic dataset, the model achieved a Cosine Similarity Accuracy (CSA) of 0.674 and a Target Registration Error (TRE) of 0.680, demonstrating superior performance.
WEI Junyi, MA Changsheng, CHEN Cheng, WANG Qian. Object Detection Technology Based on Multi-modal Models[J]. Technology of IoT&AI, 2026, 58(4): 63-68.
In the context of the rapid development of artificial intelligence technology, military target detection technology has become a key to enhancing battlefield perception and decision-making capabilities. Relying on large-scale models to conduct research on military target detection has become a new trend. A military target detection framework based on the collaboration of large models and small models is proposed. By integrating the semantic understanding ability of multi-modal large models (Qwen-VL) and the real-time detection advantages of small models (YOLOv11/YOLOv11-Pose), high-precision identification and threat assessment of targets in complex battlefield environments are achieved. Experimental results demonstrate that the large-small model collaboration framework performs remarkably on a self-built military dataset, mAP@0.5 reaches 0.82, the false positive rate drops to 15%, the end-to-end latency is kept under 500 ms, and correlation coefficient between threat scores and expert annotations is 0.89, satisfying the real-time demands of military scenarios. The research results provide technical references for the military’s intelligent transformation and verify the feasibility and application potential of the collaboration of large and small models in the target detection task.
ZHANG Houyou. Application of Artificial Intelligence Technology in Automated Monitoring of Urban Waste Incineration Power Generation[J]. Technology of IoT&AI, 2026, 58(4): 69-72.
Based on the demand for automated monitoring of urban waste incineration power generation, this paper studies the structure of the monitoring system and the key parameter characteristics during the operation of waste power generation. It elaborates on monitoring methods such as abnormal state recognition, equipment operation assessment, operation parameter optimization, and fault prediction and early warning, and introduces the application of multi-parameter data analysis in operation status judgment. Taking a certain waste incineration power plant as an example, system tests were carried out. The test results show that this method can improve the monitoring response speed and the accuracy of abnormal recognition, and achieve early warning of operation risks.
CHEN Shiyang, LIU Faming. Design of Shipborne VDR Automated Testing Device Based on Machine Vision[J]. Technology of IoT&AI, 2026, 58(4): 73-76.
In view of the insufficient degree of automation in the testing process of the Voyage Data Recorder (VDR) on board ships, a fully functional automated testing device based on machine vision is designed. The overall system architecture, the design concept of the visual inspection module and the image recognition algorithm are expounded. The implementation method of the Programmable Logic Controller (PLC) linkage control is introduced. The test results show that the system performs excellently in terms of recognition accuracy, response speed and stability, and can meet the full-process automatic detection requirements of shipborne equipment.
ZHANG Lianzhong, QI Wanxu. Multi-source Sensing Internet of Things Based Crack Control Technology for Large-volume Concrete in Cold Regions[J]. Technology of IoT&AI, 2026, 58(4): 77-80.
Based on the engineering situation in high-altitude areas where large-volume concrete is prone to cracks and their evolution is complex, this study explores the collaborative technology of crack perception, risk warning and regulation. It elaborates on the method of crack feature identification and risk determination driven by multi-source monitoring data, and introduces the linkage realization mechanism of risk results and on-site regulation measures. Combined with actual engineering application cases, the operation process and control effect of the integrated system are verified. The results show that the proposed technical system can achieve continuous monitoring and dynamic regulation of crack conditions, and has good engineering applicability.
LI Qiuyu, LI Qiyang. Optimization of Image Classification Algorithm Based on Convolutional Neural Network[J]. Technology of IoT&AI, 2026, 58(4): 81-84.
In image classification tasks in complex scenarios, traditional algorithms often have problems such as the lack of representativeness in feature extraction and the easy confusion of similar categories, resulting in unsatisfactory classification results. This paper proposes an image classification algorithm optimization scheme based on convolutional neural networks from three aspects: data set preprocessing and enhancement optimization, optimization of image classification feature extraction models, and post-processing correction of decision boundaries. The test results show that the proposed optimization algorithm can improve the accuracy of image classification, solve the problem of confusion of similar categories.
LI Lufei. Application of Multi-source Surveying Data Fusion in 3D Real-world Modeling[J]. Technology of IoT&AI, 2026, 58(4): 85-88.
To enhance the fusion efficiency and representation accuracy of multi-source surveying data in 3D real-world modeling, this study employs methods such as phased registration of heterogeneous data, point cloud-image mapping, and multi-scale fusion optimization to conduct a systematic analysis and engineering validation of the collaborative modeling process for multi-source data. The results indicate that the proposed method achieves a unified representation of geometric structure and texture information, significantly improves model accuracy and stability, and maintains good adaptability and operational consistency under complex scene conditions.
CUI Yunhao. Research on Unmanned Aerial Vehicle Trajectory Planning Based on Improved Ant Colony Algorithm[J]. Technology of IoT&AI, 2026, 58(4): 89-93.
To address the problems of slow convergence, numerous turning points, and a tendency to fall into local optima in unmanned aerial vehicle path planning under complex environments, an improved ant colony algorithm is proposed and applied to unmanned aerial vehicle path planning. To tackle the poor initial optimization capability of the traditional ant colony algorithm, the artificial potential field method is incorporated to simulate the attraction of food to ant colonies, and the initial distribution of global pheromones is updated. This guides the ant colony during the initial optimization process and reduces blindness in the early search stage. Elite ants are introduced into the traditional ant colony algorithm, with a larger moving space and more sensitive pheromone update rules assigned to them, so as to improve the optimization ability and convergence speed of the algorithm. Drawing on the idea of the A* algorithm, the heuristic function of the ant colony algorithm is revised by adding the distance between feasible nodes and the target node, strengthening the guidance for the ant colony. Simulation results show that the improved ant colony algorithm can provide higher-quality solutions in terms of path length, iteration times, and trajectory smoothness.
QUE Liman, KUANG Yu. Application Scenarios and Efficacy Analysis of AIGC in Modern Service Industry[J]. Technology of IoT&AI, 2026, 58(4): 94-97.
To enhance the service response efficiency and content generation quality of modern service industries, an Artificial Intelligence Generated Content (AIGC) driven service system architecture was constructed, and generative models, recommendation mechanisms, and reinforcement learning decision-making methods were introduced. Comparative experimental analyses were conducted for typical customer service and content recommendation scenarios. The results showed that the average response time was reduced by approximately 43%, user satisfaction increased to 91.3%, and the rate of human intervention dropped to 28.5%. The research suggests that AIGC can achieve service process automation and job structure optimization, and it is recommended to strengthen the integration of industry knowledge and the construction of model collaboration mechanisms.
LIU Qing, QI Kai, JIANG Yibo, MA Xiaoshu, ZHAO Yuxiang. Cherry Maturity Detection Method Based on Improved YOLOv11n Model[J]. Technology of IoT&AI, 2026, 58(4): 98-103.
In response to the problems of low efficiency and strong subjectivity in traditional cherry maturity detection methods, an improvement was made based on the YOLOv11n model. A composite detection model integrating - Single-Head Self-Attention (SHSA), Efficient Prompt Guided Operator (EPGO), and Convolutional Gated Linear Unit (CGLU) - was proposed and named YOLOv11n-C3k2-SEC. Experimental results show that compared with the unimproved YOLOv11n model, the improved model’s mAP@0.50 increased by 5.2 percentage points, and mAP@0.50:0.95 increased by 5.9 percentage points. The parameter quantity and computational cost showed a slight decreasing trend. The proposed model outperforms mainstream models such as Faster Region-based Convolutional Neural Network (R-CNN), YOLOv8 series, and YOLOv10 series in terms of detection accuracy, fine-grained classification, and applicability to complex scenarios. It can handle scenarios with dense fruits, occlusion, and background interference.
WANG Fang. Research on Cross-Domain Consistency Synchronization Management Mechanism for Multi-cloud Collaboration[J]. Technology of IoT&AI, 2026, 58(4): 104-108.
With the wide application of cloud computing technology, enterprises increasingly rely on multi-cloud environments for data storage and processing. In multi-cloud collaboration scenarios, how to ensure data consistency across multiple cloud platforms has become a problem to be solved. Based on the current research status, a cross-domain data synchronization mechanism is designed, which is based on the selection of consistency models, the formulation of synchronization protocols, and the guarantee of real-time performance and fault tolerance. The detailed system architecture and module design are carried out in combination with the actual needs of multi-cloud collaboration, and the effectiveness of the proposed mechanism is verified through specific example scenarios and experimental data analysis.
XU Feng, WANG Kemin, XING Yunhao, MA Hanbin, XU Huafeng. Laser-Based Conductor Diameter Monitoring System for Remote Areas Supported by Low Earth Orbit Satellite Communication[J]. Technology of IoT&AI, 2026, 58(4): 109-113.
The diameter monitoring of transmission lines in remote areas is limited by communication blind spots and sensing accuracy issues. To address this, a method for extracting the wire diameter based on laser measurement was studied. The measurement path, image recognition algorithm, and remote power supply scheme were analyzed, and the implementation methods of low-orbit satellite communication in link access, data caching, and channel anti-interference were discussed. The proposed system has the ability to operate stably in unmanned areas and provides a deployable monitoring and communication integration solution for complex terrain power grids.
FAN Lijun. Application of 5G in Low Latency Control of Vehicle-Road Coordination[J]. Technology of IoT&AI, 2026, 58(4): 114-117.
To meet the requirements of low latency and stable communication for vehicle-road coordination, this paper analyzes the limitations of traditional communication in terms of latency and reliability, and proposes a vehicle-road coordination control scheme based on 5G, covering network deployment, roadside perception, vehicle data transmission and edge execution. The test results show that 5G can significantly improve the stability and operational efficiency of vehicle-road coordination control, providing practical support for the application of related technologies.
LIU Yanan, LI Yafei, CHENG Pu, MENG Ranping, ZHAO Haoming, CUI Biao. Simulation Study on Fault-Tolerant Control Strategy of Wind Power Cyber Physical System Under Network Attacks[J]. Technology of IoT&AI, 2026, 58(4): 118-122.
To suppress the impact of network attacks on the operational stability of Wind Power Cyber-Physical System (WF-CPS), an improved non-fragile fault-tolerant control strategy based on Proportion Integration Differentiation (PID) is proposed. A model including the perception layer, network layer, control layer and physical layer is constructed to clarify the information-physical coupling attack propagation path. A PID fault-tolerant controller integrating state feedback and disturbance compensation is designed, and the quantitative indicators of non-fragility characteristics and the dynamic adjustment logic of parameters are given. A probabilistic Denial of Service (DoS) attack model is established and the parameter engineering basis is clarified. The fault-tolerant correction coefficient boundary constraint and robustness design are added. Simulation results show that the proposed strategy can effectively reduce the negative impact of DoS attacks on WF-CPS, providing technical support for the safe and stable operation of wind power systems in a network attack environment.
LIU Feng. Design of Intelligent Fire Sprinkler System Under Background of Urban Utility Tunnel[J]. Technology of IoT&AI, 2026, 58(4): 123-126.
Urban utility tunnels are underground lifeline facilities with enclosed spaces and dense pipelines. In the event of a fire, the response time is critical, and there are higher requirements for the accuracy and coordination of fire sprinkler systems. the overall design idea of intelligent fire sprinkler system is built around the operation characteristics of urban utility tunnel, combined with fire perception, multi-source information collection, sprinkler partition control, linkage response organization and redundant configuration, and the key methods are studied and applied test analysis is carried out.
JIANG Haifeng. Key Technology and Application Analysis of Intelligent Control in Electrical Engineering Automation[J]. Technology of IoT&AI, 2026, 58(4): 127-130.
In response to the demand for real-time perception and intelligent response capabilities in electrical engineering control systems, a multi-layer heterogeneous control architecture was constructed, and systematic design and integrated implementation were carried out for sensor network configuration, edge decision-making algorithms, predictive recognition models, and parameter self-tuning mechanisms. Performance verification was conducted in combination with the application case of the TBEA Hengyang Smart Factory. The research results show that the multi-model fusion recognition accuracy rate is 96.8%, the control response time is shortened to 93 ms, and energy consumption is significantly reduced, verifying the feasibility and performance advantages of this intelligent control system in practical engineering applications.
LIU Rui. Intelligent Assessment and Analysis of Power Equipment Operation Based on Multi-source Sensor Data Integration[J]. Technology of IoT&AI, 2026, 58(4): 131-135.
To address the issues of single data source and low fault identification accuracy in the operation status monitoring of power equipment, an intelligent evaluation and warning system for power equipment based on multi-source information fusion is constructed. The complementary enhancement mechanism of multi-source heterogeneous data in the feature space is deeply revealed, and the attention mechanism is introduced to dynamically weight the feature contributions of different modalities, solving the problem of information loss and weak anti-interference ability of single-source data under complex working conditions. On this basis, a cascaded innovative architecture of “spatial-temporal alignment - attention fusion - residual classification” is proposed, which is different from the existing serially stacked deep learning models, achieving end-to-end collaborative optimization of feature extraction and fault discrimination. A fault evolution trajectory tracking technology based on particle filtering is developed to accurately predict the remaining life of equipment and conduct scientific maintenance. To verify the effectiveness and reproducibility of the designed method, a standardized test set containing various typical faults is constructed, using a five-fold cross-validation strategy, and the performance of multi-model identification under different working conditions and the operational key indicators before and after system application are compared. System integration verification results show that the research results will significantly enhance the robustness of state identification under complex working conditions, providing technical support for the intelligent operation management of power grids.
TAO Chuanwen. Application of RFID Technology in Intelligent Warehouse Inventory Management#br#[J]. Technology of IoT&AI, 2026, 58(4): 136-139.
To enhance the efficiency of warehouse management, an intelligent inventory management method based on Radio Frequency Identification (RFID) is proposed. By establishing a data collection framework integrating RFID and the Internet of Things, real-time acquisition and update of inventory information can be achieved. On this basis, the K-means algorithm is utilized to classify the access frequency of items, thereby constructing a dynamic inventory optimization model for optimizing the warehouse layout. Experimental results show that the proposed method can effectively improve the accuracy of inventory data, shorten the processing time for entry and exit, and has good practical application value in intelligent warehouse management.
ZHOU Yafeng. Design of Internet of Things Multi-source Data Fusion and Collaborative Management Platform for Smart Ports[J]. Technology of IoT&AI, 2026, 58(4): 140-143.
The operation intensity of smart ports is continuously increasing. The concurrent operation of multiple systems and the high-frequency perception of equipment jointly contribute to the characteristics of diverse data protocols, frequency differences, and mixed structures. To meet the collaborative requirements of production scheduling and equipment maintenance, a multi-source data fusion and collaborative management platform is constructed. Relying on a unified data model and spatio-temporal alignment mechanism, sensor data, video information, and scheduling records are integrated to form a unified business view. The platform is verified in the context of automated terminals. The operational verification results show that the constructed platform exhibits better performance in both fusion processing and scheduling response, and data consistency remains stable during continuous operation, further demonstrating the feasibility of engineering applications.
LIU Yuan. Safety Risk Assessment and Optimization of Energy Enterprises Under Environmental Constraints#br#
#br#[J]. Technology of IoT&AI, 2026, 58(4): 144-148.
To meet the safety operation requirements of energy enterprises under environmental constraints, a dynamic safety risk assessment and collaborative optimization method is constructed. A risk assessment index system composed of environmental factors, equipment status factors and management factors is established. The dynamic risk assessment is realized by introducing variable weight analytic hierarchy process and cloud model, and a risk level determination method combining environmental pressure index and comprehensive risk value is constructed. On this basis, a multi-objective collaborative optimization model and dynamic regulation strategy library are constructed to achieve the collaborative control of safety risks and environmental constraints, providing a technical path for the safety operation management of energy enterprises.
ZHANG Xiyuan, ZHU Guoqiang, DU Tianyun, CHEN Chicheng. Research on Redundancy Transformation Technology and Adaptation Application of PSL Terminals in Small Platform Rail Transit Stations#br#
#br#
[J]. Technology of IoT&AI, 2026, 58(4): 149-152.
Taking the small platform of Wenzhou Rail Transit Line S2 as the research object, a comparative analysis was conducted on indicators such as call setup time, the number of physical terminal faults, communication link availability, fault detection time, and fault handling time before and after the transformation. The results show that after the redundancy transformation, the average call setup time of the Platform Screen Door Local Control Panel (PSL) terminal at the small platform is shortened, the physical fault rate of the equipment is reduced, the communication link availability is improved, and the response speed of the background operation and maintenance is accelerated. This method is suitable for the small platforms of existing lines, lightweight communication transformation, and terminal upgrades while retaining the main framework of the original system, which can enhance the reliability of the communication system of the small platforms in rail transit.
YU Shijun, LI Kun, HE Xiaoxian, WANG Yunfeng, ZHANG Hongbo. Design and Implementation of Satellite Earth Station Control and Management System Based on B/S Architecture#br#
#br#
[J]. Technology of IoT&AI, 2026, 58(4): 153-157.
In response to the problems of poor compatibility and insufficient expansion capability of the satellite earth station station control management system, a satellite earth station control management system based on the Browser/Server (B/S) architecture was designed and implemented. The basic composition, external interfaces, working process, core functions and software implementation of the system were elaborated in detail. This system uses the Transmission Control Protocol (TCP) to achieve unified control of station equipment, is simple to deploy, has concise and intuitive software pages, and has flexible device expansion capabilities.
REN Xinbo, QI Jianghao. Design and Application of Remote Monitoring System for LNG Loading Management Based on PLC Technology[J]. Technology of IoT&AI, 2026, 58(4): 158-162.
In the remote monitoring process of Liquefied Natural Gas (LNG) loading management, analog circuit control technology is commonly employed to achieve remote loading control. However, this technology is susceptible to interference from external electromagnetic fields, leading to inaccurate control results. Therefore, this study proposes a Programmable Logic Controller (PLC)-based remote monitoring system for LNG loading management. In terms of hardware, the system uses a PLC controller as the core, integrating batch control instruments, weighing scales, computers, and other devices to form the hardware architecture of the remote monitoring system. In terms of software, the system first applies evidence theory to fuse data from multiple weighing scales, enabling remote dynamic monitoring of LNG loading weight. Secondly, an optimization decision-making objective function for LNG loading management is constructed, and the particle swarm optimization algorithm is used to determine the optimal loading management plan. Finally, the fuzzy Proportion Integration Differentiation (PID) control algorithm is employed to execute loading management decisions, achieving remote fuzzy control for LNG loading. Application results demonstrate that the average control error of the system is 1.24%, the loading efficiency reaches 28.8 t/h, the system average response time is shortened to 287 ms, and the loading qualification rate reaches 97.5%, effectively improving the control accuracy and comprehensive performance of LNG loading management.
WANG Zhiwei. Construction and Application of Digital Information Management System for Hydropower Station[J]. Technology of IoT&AI, 2026, 58(4): 163-167.
A three-layer Browser/Server (B/S) architecture design for an electric power station digital information management system based on SpringCloud microservices and Vue.js framework is proposed. The system is divided into three core business modules: basic information management, Geographic Information System (GIS) information display, and digital management results. It integrates multi-terminal real-time communication, distributed real-time computing, and an integrated visualization technology support system for operation. The system performance was verified through a 30-day pilot test in a small and medium-sized hydropower station. The results show that the response time of the proposed system under 100 concurrent users is less than 400 ms, and the cross-terminal synchronization accuracy and overall function coverage are high, which can meet the digital management requirements of the entire process of the hydropower station.
LI Li, WEI Jiubo, WU Zhixian. Intelligent Recognition System for Hard Hat Wear Detecting in Underground Coal Mines Based on AI Vision Technology[J]. Technology of IoT&AI, 2026, 58(4): 168-173.
To enhance the real-time monitoring and recognition stability of hard hat wear in underground coal mines, this study employs an edge-computing-based visual recognition method. It investigates target detection and wear status determination processes under complex working conditions, constructs a coordinated system integrating image acquisition, data processing, and multi-scale detection models, and designs an anomaly alarm linkage mechanism. This method maintains high recognition accuracy and low response latency in low-light and occlusion environments, demonstrating excellent on-site adaptability and operational stability.
LI Jing. Research on Smart Operation and Maintenance Mode of Unattended Pipeline Stations Based on Internet of Things#br#
#br#
[J]. Technology of IoT&AI, 2026, 58(4): 173-176.
Addressing the issues of data intermittency, delayed anomaly identification, and insufficient remote operation coordination in the operation and maintenance of unmanned or sparsely manned long-distance pipeline stations, this paper introduces an Internet of Things perception access method under existing measurement and control conditions. It designs a smart operation and maintenance mode that includes measurement and control data aggregation access, dynamic continuous collection of operational parameters, quantitative diagnosis of equipment conditions, and closed-loop execution of remote control. This mode achieves data-driven station operation and maintenance through dynamic sampling, deviation rate calculation, instruction logic verification, and other technologies. Comparative tests conducted based on actual oilfield stations show that this mode significantly improves data continuous coverage, shortens anomaly response and remote control response times, and increases the success rate of instruction execution. This mode promotes the transformation of station operation and maintenance from a patrol-oriented approach to a data-driven one, providing a practical engineering reference for optimizing the operation and maintenance of unmanned long-distance pipeline stations.
