memory
SERIA C
Inginerie Electrică și Informatică
Excelență în Cercetare
Din 1929
Arhiva Online
Căutare
Ani
DESIGN AND APPLICATION OF HUMAN-COMPUTER COLLABORATIVE TEACHING INTERACTIVE SYSTEM DRIVEN BY ARTIFICIAL INTELLIGENCE
person
Zhijuan WANG
It is challenging for teachers to quickly and accurately assess each student's learning status due to the ongoing growth in classroom size and the complexity of the curriculum in contemporary educational settings. This limits classroom interaction and the efficacy of individualized instruction....
INTEGRATING BOUNDARY AWARENESS AND DYNAMIC LOSS FOR ENHANCING SEMI-SUPERVISED NER IN POWER COMMUNICATION FAULT TEXTS
person
Jieke LU, Ying LING, Xin LI, Shaofeng MING
As an important data source for troubleshooting, power communication fault text holds significant value in fault diagnosis. However, the effectiveness of generic named entity recognition (NER) techniques is limited due to the irregular nature of domain-specific text, often characterized by small...
STRATEGIES FOR DEVELOPING THE ICT ECOSYSTEM IN THE REPUBLIC OF MOLDOVA
person
Lilia SAVA, Valentina TIRSU, Dinu TURCANU, Elena PETREANU
In the abstract, the authors examine how the Republic of Moldova can strengthen its ICT ecosystem by leveraging AI, IoT, and 5G to modernize digital infrastructure and accelerate innovation. Using a mixed approach that combines policy review, analysis of official connectivity and sector...
RESEARCH ON VIRUS PROPAGATION MODELS IN CLOUD NETWORKS
person
Yanjing LI
Traditional virus propagation models exhibit notable limitations when describing viral spread in cloud networks, particularly in their passive treatment of external factors, ignorance of network dynamics, and oversimplified immunization strategies. To address these gaps, this paper proposes an...
APPLICATION OF ANT COLONY OPTIMIZATION ALGORITHM BASED ON CLOUD MODEL IN ROUTING PROBLEMS
person
Xu LI, Zhengyan LIU
A cloud model-based ant colony optimization algorithm was proposed in this research to address the problem of difficult balance between exploring the search domain and developing known solutions in ant colony optimization algorithm. Deep analysis of the characteristics of cloud models could...
SECONDARY FREQUENCY REGULATION CHARACTERISTIC MODELING OF THERMAL POWER UNITS BASED ON BOUNDED-GAIN-FORGETTING
person
METHOD, Xiufeng XING, Fengsheng JIA, Xinglei YAN, Jian LV, Honghai LI
The existing mechanism models for thermal power units face challenges in accurately reflecting unit response characteristics under different working conditions. To address this, this paper proposes a novel modeling method for the secondary frequency regulation characteristics of thermal power...
DISTRACTED DRIVING BEHAVIOR DETECTION BASED ON CCLA-YOLOv8
person
Weihua LIU, Lele LV, Wei SHENG
With the acceleration of urbanization and the continuous growth of vehicle ownership, traffic safety issues caused by distracted driving behaviors have become increasingly prominent. To address this problem, this paper focuses on the research and design of a driver behavior detection system...
EXPLAINABLE ARTIFICIAL INTELLIGENCE MODELLING OF A COMBINED CYCLE POWER PLANT USING KOLMOGOROV-ARNOLD NETWORKS
person
Paul-Vasile VEZETEU, Dumitru-Iulian NĂSTAC
Artificial Intelligence Explicability (XAI) is an essential research field which aims at developing methods and models which can explain the decision process of machine learning (ML). Interpretable ML is a subfield concerned with developing algorithms which are intrinsically explainable. The...
SECURITY-AWARE UNIVERSITY RESOURCE ALLOCATION: ECC-BASED VERIFICATION FOR DYNAMIC COORDINATION
person
Xiaomin ZHAO
The allocation of educational resources in higher education institutions faces challenges related to execution efficiency, coordination stability, and data security under dynamic demands and multi-stakeholder participation. This study proposes a security-aware dynamic allocation framework for...
RESEARCH ON COOPERATIVE PASSIVE LOCALIZATION FOR UAV FORMATIONS BASED ON GRAPH MODEL AND DISTRIBUTED A* SEARCH
person
Liqiong PAN, Jiajia QIAN, Haixu ZHANG
Bearing-only passive positioning is the key to achieving autonomous collaboration of UAV swarms in GPS-denied environments. This paper proposes a two-layer algorithm framework that integrates graph models and distributed search to solve this problem. In the first stage, a fast topological...
MEDICAL IMAGE SEGMENTATION BASED ON SPATIAL INFORMATION ENTROPY GEODESIC ACTIVE CONTOUR
person
MODEL, Xiang MA, Zhimin WANG, Haibo ZHANG, Ronghao ZHANG
This paper presents a novel Geodesic Active Contour (GAC) model based on spatial information entropy for the medical images including bias fields and weak edges. The driving term of this model includes two parts: This spatial information entropy based on image can well describe the gray...
WHALE OPTIMIZATION ALGORITHM BASED ON TENT AND SINE OPTIMIZATION
person
Jing CHEN
To address the limitations of the whale optimization algorithm (WOA), including its tendency to fall into local optima and weak local search capability, an improved version—termed TSVWOA—is proposed in this paper. Specifically, the tent map is adopted for population initialization to enhance...
ON-AXIS TREATMENT METHOD OF THE BOR-FDTD ALGORITHM AND ITS APPLICATIONS
person
Miaomiao LI, Jie JI, Shanshan SHEN
Originating from the standard Finite-Difference Time-Domain (FDTD) method, the Body of Revolution FDTD (BOR-FDTD) algorithm stands out as a high- efficiency computational paradigm, dedicated to simulating the propagation behaviors of electromagnetic waves in structures with pure symmetry, e.g.,...
DESIGN OF AN INTEGRATED INTERACTIVE EXPERIMENTAL SYSTEM FOR TEACHING DEEP LEARNING-ORIENTED OBJECT DETECTION
person
Lili ZHANG, Wenshuo HAN, Yao HUANG, Hongxin TAN, Wei WEI, Jing, LI, Pei YU, Lingyu ZHANG
Artificial intelligence has become an important part of university education worldwide. Deep learning courses are offered in many universities as core courses in the field of artificial intelligence. To enable students to comprehensively understand and master the knowledge system and content of...
RESEARCH ON 5G BASE STATION ENERGY EFFICIENCY BASELINE CONSTRUCTION TECHNIQUE BASED ON PRINCIPAL COMPONENT ANALYSIS AND ARTIFICIAL NEURAL NETWORKS IN THE PERSPECTIVE OF HIGH- DIMENSIONAL DATA ANALYSIS
person
Yao RAO, Yuan CHENG, Xinlei BAI, Sheng DING, Shen XIAO, Ziyi, LIU
Aiming at the high-dimensional characteristics of the data presented in the process of energy efficiency assessment of 5G base stations, this paper proposes a novel baseline construction method for energy efficiency of 5G base stations by organically combining the principal component analysis...
DAD-YOLO: EFFICIENT FIGHTING IMAGE RECOGNITION
person
MODEL
Zheng LI1,2, Dong WANG1,2,*, Dengcong MU1,2,* In today's society, fight detection technology is vital for preventing violence. To address issues like low accuracy, slow speed, and high complexity in existing video fight detection algorithms, we propose DAD-YOLO, an efficient model. Experiments...
INTELLIGENT QUESTION-ANSWERING SYSTEM FOR GRID SCHEDULING USING KNOWLEDGE GRAPH AND LARGE LANGUAGE MODEL
person
Wei LI, Bo ZHOU, Siqi SHEN, Zhangguo CHEN
With the increasing complexity of power system operations, scheduling tasks require more advanced knowledge acquisition and question-answering systems. Existing power grid scheduling question-answering systems have issues such as complex operations and incomplete management frameworks. To...
RESEARCH AND DESIGN OF MULTI-MODAL FUSION + ATTENTION-TCN EMOTION RECOGNITION ALGORITHM BASED ON RASPBERRY PI
person
Li WANG, Xin WANG, Yandong LIU, Chao CHEN, Lei TANG
To address the limitations of traditional emotion recognition algorithms— reliance on single-modal expressions, weak temporal modeling, susceptibility to spoofing, and poor adaptability on embedded devices like Raspberry Pi—this study optimizes emotion recognition technology for Raspberry Pi....
MACHINE LEARNING (ML) BASED MITIGATION OF HARMONICS AND POWER FACTOR IMPROVEMENT
person
Kishore Kumar Pedapenki, Suresh Kumar Annam, G. Pandu Ranga Reddy
Power Electronics is a branch of Electrical Engineering which has enor- mous applications in changing the one type of supply into another and also obtain- ing the different levels of voltages and current based on the requirement. Such an important branch also creates some problems if the same is...
OPTIMIZING EFFICIENCY IN MULTI-STAGE RESONANT WPT BY CO-DESIGNING COIL INNER RADIUS AND TURN
person
COUNT, Gang LIANG, Xin WANG, Chunlong LI, Yongbo YAO, Qing LI
For medium- and long-range contactless power delivery, multi-stage resonant magnetic wireless power transmission (WPT) is a compelling option; however, transmission efficiency remains the central hurdle to practical deployment. This paper aims to improve the efficiency of multi-stage magnetic...
ENHANCED ADAPTIVE LOW-RANK ADAPTATION WITH RANK ADJUSTMENT AND MODULAR FUSION FOR EFFICIENT FINE-TUNING
person
Xiangjiang LI, Tiejun PAN, Diqun YAN, Zheqi ZHANG, Enhui HU, Junyi CHAI, Samuel Ken-En GAN, Yuan XU, Hongbo LI
Low-rank adaptation (LaRA) is an efficient approach for fine-tuning pre- trained models, leveraging low-rank matrix decomposition; prior research on LoRA has investigated a range of model optimization techniques, where low-rank matrix adaptation is employed to effectively optimize pre-trained...
CASE STUDY FOR HYDROGEN PRODUCTION USING A MODULATED CURRENT SOURCE
person
Viorel Constantin PETRE, Octavian Mihai GHITA, Cerasela DINU
Hydrogen is a promising alternative fuel that can offer a range of benefits over traditional fossil fuel. The common element, water, is seen as the best renewable and non-polluting fuel on the planet. The article proposes and presents an analysis of water electrolysis using a PWM modulated...
AN INDUSTRIAL AUTOMATION ARCHITECTURE FOR A DEBURRING CELL WITH ROTARY INDEXING TABLE
person
Florian-Daniel DOGARU, Violeta-Georgiana DOGARU, Valentin, NĂVRĂPESCU, Alexie GHERGHE
This paper presents the design and implementation of an industrial automated equipment used for deburring parts produced for the automotive industry. The system contains a rotary table with several machining stations in a PLC-controlled environment. The proposal consists of electrical design and...
CURRENT STATE OF ANTSIRANANA'S ELECTRICITY NETWORK AND PROPOSED SOLUTIONS
person
M.M. Inès MANAMIHANTA , Dominique G.M.B. RAKOTO , Liva F., RAFANOTSIMIVA, Nirinarison Jean RAZAFINJAKA, Virgil DUMBRAVĂ
This article describes the current state of the Antsiranana power grid, incorporating a photovoltaic power plant. Unlike in industrialized countries, where grids are interconnected and can be considered to have infinite capacity, those in developing countries are often standalone and have...
DESIGN AND DEVELOPMENT OF A COMPACT INTELLIGENT CABLE PULLING SYSTEM FOR POWER
person
CABLES, Yan FENG, Yufang LIN, Zongsheng XU, Sikai XIONG, Xinyu PEI, Xiongfeng HUANG
With the continuous expansion of urban power networks and increasing complexity of underground pipeline systems, traditional cable traction methods are limited in construction safety, l intelligence, and operational efficiency. To address these challenges, a compact intelligent traction system...
FAULT DIAGNOSIS AND DECISION SUPPORT SYSTEM FOR POWER GRIDS BASED ON DATA AND KNOWLEDGE
person
Xiaogang WU, Qingbiao LIN, Shengmin QIU, Yan GUO, Xingwang CHEN, Zhizhong LI
The decision-making process for power grid faults involves multiple dimensions of information. Effectively integrating this information and achieving intelligent optimization are essential for improving power grid security. Traditional fault diagnosis relies on a single data source and often...
MULTITEMPORAL DATA FUSION AND REASONING IN POWER GRID FAULT DECISION SUPPORT
person
Yipu ZHANG, Qingbiao LIN, Shengmin QIU, Kun ZHANG, Xiaogang, WU, Xin GAO
As the scale of power grids continues to expand, the volume of grid-related information increases significantly. Diagnosing the operational state of power systems by integrating real-time data streams with historical fault data has become a key focus in the field of electric power. However,...
CONFIDENCE ANALYSIS AND DEEP LEARNING-BASED APPROACH FOR TWO-STAGE CALIBRATION OF ERROR- DOMINANT PARAMETERS IN POWER SYSTEM SIMULATION
person
Shengliang CAI, Guobin FU, Mingzhi CHEN, Guoqiang LU, Haibin SUN, Xuebin WANG, Rui SONG, Yanzhe CHENG
Accurate identification and calibration of error-dominant parameters (EDPs) are critical for reliable power-system simulations. This paper presents a deep- learning-based framework for unified EDP identification and calibration. A large synthetic dataset is constructed by perturbing model...
ASSESSMENT OF THE POWER QUALITY IN THE CIRCUIT OF A STORAGE SYSTEM. EXPERIMENTAL STUDY
person
George Alexandru FLOREA, Ciprian PREDA, Octavian ANDRIȚOIU
The study includes an experimental analysis of the power quality in an industrial installation in which a solar source supplies the user and determines, during hours of excess energy, the charging of a battery and during peak hours it contributes to reducing the energy taken from the public...
INTELLIGENT GENERATION AND VERIFICATION OF START-UP SCHEMES BASED ON POWER GRID
person
TOPOLOGY, Yaopu ZHENG, Xiongfeng LIN, Shifeng PENG, Guangyu DENG, Jianguo, YE, Zhou HUANG
With the continuous expansion of the power grid scale, the traditional power grid startup schemes are facing challenges in terms of efficiency and accuracy. This paper proposes an intelligent startup scheme generation and verification technology based on power grid topology. By constructing a...