AEEE. 2025, vol. 23

Permanent URI for this collectionhttp://hdl.handle.net/10084/158014

Browse

Now showing 1 - 20 out of 26 results

Mutual Influence AI: Trust-Based Cooperation Mechanisms for LLM Multi-Agent Systems
(Vysoká škola báňská - Technická univerzita Ostrava, 2025) Oujazský, Václav; Novák, Pavel
This paper introduces Mutual Influence AI, a novel concept for adaptive cooperation in multi-agent systems. Unlike classical independent reasoning or cen- tralized orchestration, our approach introduces an ex- plicit mutual influence factor μ that captures trust- adjusted peer feedback and directly modulates large lan- guage model (LLM) generation. We present (i) a math- ematical formalization of mutual influence, (ii) a pro- totype implementation integrated with Microsoft Auto- Gen for LLM-based agents, and (iii) qualitative evi- dence that the framework improves adaptability, trans- parency, and coordination in multi-agent dialogues. Results show that Mutual Influence AI stabilizes group interactions efficiently while providing interpretable control over how agents influence each other. This positions Mutual Influence AI as a new paradigm for LLM-driven multi-agent systems with potential appli- cations ranging from collaborative problem solving to cybersecurity. Quantitatively, across 167 simulation runs, cross–role agreement increased from 0.19 (base- line) to 0.50 under influence (approx. +160%), with median revision depth (approx. 1.0). Under adversarial feedback, agreement still improved (0.18 to 0.47).
ABO-BTI: An Open-Source ABO Blood Typing Image Dataset for Medical AI Applications
(Vysoká škola báňská - Technická univerzita Ostrava, 2025) Sara, Daas; Hatem, Zehir; Asma, Chebli; Toufik, Hafs; Chaima, Hadef
Accurate blood type classification is cru- cial for safe transfusions and clinical decision-making, yet existing research is limited by the lack of stan- dardized, publicly available datasets for training and evaluating machine learning models. To address this gap, we introduce ABO-BTI (ABO Blood Typing Im- age), the first open-source dataset dedicated to blood type classification using high-resolution agglutination images. The dataset comprises 144 cases, with 432 images standardized to a resolution of 1280×590 pix- els after processing. This study evaluates the effective- ness of deep learning for blood type identification us- ing the ABO-BTI database. Three models, ResNet50, MobileNetV2, and a proposed deep learning architec- ture, were trained and tested on the dataset to as- sess its suitability for machine learning applications. The proposed model achieved an accuracy of 96.51%, significantly outperforming MobileNetV2 (12.64%) and ResNet50 (72.41%). Comparative analysis with tradi- tional machine learning methods further demonstrated that deep learning provides competitive performance while reducing reliance on handcrafted feature extrac- tion. These results highlight ABO-BTI as a valuable benchmark for advancing AI-driven blood type classifi- cation. The findings also suggest the potential integra- tion of deep learning-based classification into embed- ded systems for real-time blood typing in point of care and emergency settings. By providing a standardized dataset and demonstrating the viability of deep learn- ing models, this study lays the foundation for future re- search in automated blood classification, with implica- tions for both clinical applications and AI-driven med- ical diagnostics.
Unmanned Aerial Vehicles (UAV) in Cellular Network: Approach, Architecture, and Challenges
(Vysoká škola báňská - Technická univerzita Ostrava, 2025) Ibrahim, N. S. Amani; Asraf Saparudin, Faiz
Unmanned Aerial Vehicles (UAVs) have emerged as critical components of modern cellular net- works, offering potential for significantly enhancing cov- erage, capacity, and connectivity. This paper investi- gates the novel integration of UAVs into existing cellular infrastructure, presenting a comprehensive framework for their deployment and management. The study delves into the architectural complexities of UAV-cellular net- works, emphasizing design considerations and techno- logical requirements for seamless integration. Potential benefits, including rapid network deployment in remote areas, disaster recovery, and dynamic network optimiza- tion, are explored. However, the integration of UAVs also introduces challenges related to resource allocation, channel modeling, optimal placement and trajectory, and interference management. This paper provides a thorough analysis of these challenges, offering poten- tial mitigation strategies and innovative solutions. As a contribution to future research, a novel approach is suggested for optimizing UAV placement and trajectory to enhance network coverage and efficiency in serving ground Internet of Things (IoT) devices. Through a holistic examination of UAVs in cellular networks, this study offers a comprehensive overview of approaches, architectures, challenges, and optimization techniques for effective UAV integration.
Application of CSA Algorithm for the PMSM Speed Estimator of The FOC Control Method Using Extended Kalman Filter
(Vysoká škola báňská - Technická univerzita Ostrava, 2025) Tran, Thinh Cong; Brandštetter, Pavel; De, Huynh Tan; Vo, Hau Huu
Nowadays, Permanent Magnet Syn- chronous Motors (PMSM) are used more and more widely due to their advantages over other types of motors, such as high efficiency, constant torque, higher power density, and wide speed range. Many studies on this motor have been carried out in the industry. This paper proposes an application for the PMSM motor to estimate the speed of the motor rotor using an extended Kalman filter (EKF). This also means that the motor is controlled without using a speed sensor, so the system has the advantages of reducing the cost of manufacturing encoders, less damage, increased reliability, and reduced size due to the absence of moving mechanical parts of the sensor. The estimated performance depends heavily on the parameters of the covariance matrices in the filter. In the paper, the filter parameters are optimized using the Cuckoo Search Algorithm (CSA). The simulation results of the proposed algorithm on the PMSM motor show its advantages over traditional methods.
Exploring the Applications of Textural Features for Automatic Leather Characterization
(Vysoká škola báňská - Technická univerzita Ostrava, 2025) Skrobo, Esma; Sokic, Emir
This paper examines the significance of tex- tural features in digital images of leather samples for automated visual quality inspection in industrial produc- tion. Leather defect detection involves tackling two im- portant challenges: first, accurately isolating the leather surface from the background in the acquired images, and second, conducting a detailed analysis of the extracted region to identify and classify potential defects. This study investigates the potential of textural descriptors for leather characterization, exploring their application as feature vectors in both supervised and unsupervised machine learning methods. We evaluate these meth- ods on two tasks: distinguishing between the leather surface and background in acquired images, and classi- fying leather defects. As anticipated, supervised methods demonstrate superior performance, achieving over 98% accuracy in leather-background separation and up to 90% in defect classification. In contrast, the unsupervised approach yields more modest results, with Rand Index and Fowlkes-Mallows Index values of 81% and 73%, respectively. Despite the limitations of textural descrip- tors in leather defect classification, the results highlight the potential of texture analysis and unsupervised learn- ing in automating image analysis and enhancing quality control in leather production.
Implementation and Control System Development of a Differential Drive Wheeled Mobile Robot
(Vysoká škola báňská - Technická univerzita Ostrava, 2025) Vu, Tri-Vien; Nguyen, Thanh-Quang; Duc, Anh-Minh Tran; Nguyen, Phuong-Uyen Le; Trinh, Chan-Kiet
his study presents a cost-effective differen- tial wheeled mobile robot system designed for educational and research applications. The robot employs an ESP32 microcontroller for control and communication, integrat- ing sensors such as ultrasonic modules, an IMU 9050, and a line-following module via the I2C protocol. PID controllers were designed and fine-tuned using MAT- LAB’s System Identification Toolbox to address motor asymmetries, ensuring stable and synchronized perfor- mance. A custom-built user interface developed in C# enables real-time monitoring and control via Bluetooth, allowing users to configure modes and visualize robot trajectory. The experimental setup was intentionally simple and reproducible, enabling straightforward de- ployment for instructional purposes. The system was validated through tasks such as line-following, obsta- cle avoidance, and trajectory tracking, with the latter demonstrating superior accuracy. While limitations, such as sensor inaccuracies in reflective environments and computational constraints of the ESP32, were iden- tified, proposed future enhancements include integrat- ing advanced sensors and machine learning algorithms. The robot’s modularity, affordability, and adaptability make it a versatile platform for introducing students to robotics and conducting foundational research in control systems.
Investigation of Modified New Hybrid Multilevel Inverter Topology with Reduced Switches
(Vysoká škola báňská - Technická univerzita Ostrava, 2025) Manivel, Murugesan; Tamilselven, Kesavan; Nagulsamy, Suganya; Palani, Lakshmanan; Samarasam, Brindha; Karuppannan, Anand; Rajaduraipandi, Jayanthi; Subramanian, Balambigai
In this article, a modi ed new hybrid mul- tilevel inverter is suggested. This system consider- ably diminishes the switches to twenty-seven to pro- duce fteen-level, three-phase output, whereas a tradi- tional multilevel inverter requires more semiconductor switches. The projected multilevel inverter considerably crops better response and stretches high e ciency of 98.35% owing to less switching and conduction losses, reduced cost, improved power quality, and high reliabil- ity. The nearest level control scheme is used for con- trolling the switching devices. Due to the lower counts of switches, complexity of the switching circuitry is less. The total standing voltage across the switches is 40Vdc and the cost function per unit is 5.714 for a weight co- e cient of 1, which is also less when compared to the traditional multilevel inverters. Total harmonic distor- tion produced by this inverter is 4.53%, which is under the IEEE standard of 5%. So, the suggested multilevel inverter is best-suitable for electric vehicles.
Forecasting of Rainfall and Identification of Its Scarcity in Indian States for the Upcoming Decade
(Vysoká škola báňská - Technická univerzita Ostrava, 2025) Sharma, Leena; Verma, Gauran
This research delves into the analysis of rain- fall data for the past 50 years in India to identify criti- cal areas necessitating the implementation of rainwater harvesting systems. The impetus for this investigation arises from the rising demand for such systems, primar- ily driven by dwindling groundwater levels in certain Indian regions. These declining groundwater levels can be attributed to reduced rainfall, which, in turn, can be linked to the effects of global warming and the unpre- dictability of weather conditions. The study employs time series analysis techniques, particularly utilizing the neural prophet model, to forecast the rainfall in the upcoming decade for the different states in India. Based on the forecasting data, the most significant deficits in terms of annual rainfall are identified using the percent- age departure formula available on the IMD (Indian Meteorological Department) website. The result shows that the following states are the most deficient, i.e., Mathatwada region, and Haryana Chandigarh region, in terms of scarcity of rainfall. It has been observed from the analysis that the error is in the range of ±1%. The predicted rainfall in Haryana, Chandigarh is -683.53 mm, and Mathatwada is 5865 mm
MULTI-OBJECTIVE INTEGRATED APPROACH FOR DISTRIBUTED GENERATION PLACEMENT AND SIZING TO ENHANCE PERFORMANCE OF RADIAL POWER DISTRIBUTION SYSTEM
(Vysoká škola báňská - Technická univerzita Ostrava, 2025) Ranga, Jarabala; Jalli, Ravi Kuma; Muthusamy, Senthil Kumar; Sundaresan, Vijayalakshmi; Ramesh, Sundar; Palanisamy, Rajakumar; Aralprakusam, Sakthidasan
This article proposes an integrated approach combining a novel population-based jellyfish search al- gorithm (JSA) and loss sensitivity factor (LSF) to optimize single and multiple distributed generation (DG) placement and ratings for total active power loss (TAPL) minimization and total voltage deviation (TVD) reduction. The study considers photovoltaic (PV) and wind turbine (WT) DG systems for optimal integration. The simulation findings are examined for a single and multiple DG allocation on the IEEE 69-bus benchmark radial distribution power network (RDPN). The TAPL of the 69-bus benchmark RDPN is mini- mized from 225 kW to 77.10 kW, 62.32 kW, 19.52 kW, and 9.94 kW after the single PV, three PV, single WT, and three WT DG systems optimization, respectively. Simultaneously, the TVD is minimized from 1.8369 per unit (p.u.) to 0.7036 p.u. and 0.6698 p.u. for a single and three PV DG systems optimization, respectively, and 0.3934 p.u. and 0.3466 p.u. for a single and three WT DG systems placement, respectively. The perfor- mance of the proposed integrated approach is compared to the different optimization techniques, taking TAPL as a comparison metric. The comparison showcases that the integrated approach results in a favorable op- timal solution among the compared optimization tech- niques.
A Novel Method for Multiple Sound Sources Localization with Low Complexity
(Vysoká škola báňská - Technická univerzita Ostrava, 2025) Hieu, Nguyen Trung; Yamada, Kou
Sound source localization is essential in many areas such as robotics interaction, teleconferenc- ing, sound extraction and recognition, noise cancellation in vehicles, object location detection, assessment of noise pollution in living spaces, and search and rescue. Inter- action in natural settings requires the detection of differ- ent sources of sounds from the environment. Accurately detecting and differentiating incoming sound directions always attracts attention and has been researched us- ing various methods. However, most of these methods still require complex algorithms or large amounts of calculations, which are accompanied by the cost of hard- ware and system resources. In this paper, we present a novel method and metrics for estimating the direction of multiple sound sources based on a combination of beamforming, time difference of arrival (TDOA), and frequency sparsity. Our new proposals are well-suited for deployment on resource-limited devices, offering re- duced processing complexity, short computation time, and real-time response.
Bearing Fault Diagnosis Based on Convolutional Neural Network using Estimated Motor Current Signals and their Spectral Portrait
(Vysoká škola báňská - Technická univerzita Ostrava, 2025) Dang, Huu Hai; Bui, Ngoc-My; Hoang, Van-Phuc; Bui, Quy Thang; Doan, Van Sang
Induction motor bearing fault diagnosis stands as a crucial aspect of rotating machinery maintenance. Numerous studies have delved into employing current signals and machine learning methods for this purpose. However, the effectiveness of these approaches relied heavily on manually selecting features for training. Moreover, traditional machine learning techniques struggle with large volumes of computational data. To address these limitations, researchers have turned to deep learning architectures such as Convolutional Neural Networks, ResNet, and AlexNet, either individually or in combination with traditional machine learning methods, for bearing fault diagnosis. Published convolutional neural network-based works usually use basic CNN networks. The experimental data are time or frequency domain data, and the fault classification accuracy is high only with noise-free signals. This paper proposes a novel approach aimed to enhance the accuracy of bearing fault identification by leveraging a CNN model trained on both the estimated motor current signals and their corresponding Fast Fourier Transform values. Comparative analysis against existing methodologies including machine learning and single-input convolutional neural networks or multiinput convolutional neural networks demonstrates that the proposed method achieves impressive results. The bearing fault accuracy reaches up to 99.88% for noisefree signals and 99.14% for signals with added noise at a Signal-to-Noise Ratio of -10 dB.
A Design of Smart Feeding System for Dairy Cows Farm
(Vysoká škola báňská - Technická univerzita Ostrava, 2025) Tran, Nhat Minh; Tran, Thang Viet; Nguyen, Ngo Minh Tri; Ho, Minh Nhut; Nguyen, Chi-Ngon
In this study, we present the design of an automatic food supply system for dairy cows using an expert calculation formula based on collected data, including temperature and humidity of the barns, cow movement, daily milk production, the breeds and ages of the herd. When the expert system calculates the amount of the total mixed ration (TMR) feed to provide to the cows, it will transfer the TMR value to the proposed system to provide the exact amount of the daily feed. The designed feeding system performs 3 tasks: 1) Accurately quantify the amount of the TMR feed ingredients according to the formula including chopped grass and fine feed portions; 2) Mixing the measured ingredients during a preset time from the system; 3) Transporting the feed after mixing to the feeding trough for dairy cows and then automatically retrieving the leftover feed. The proposed system was tested at the Tan Tai Loc dairy cow farm, Soc Trang City, Vietnam. The results show that the automatic weighing system can reach ±0.5% accuracy, and the proposed feeding system can minimize feed loss and eliminate overfeeding to promote good cow health and performance.
Zebra Optimization Algorithm for Power Conditioner with Fractional order PID Controller for Power Quality Improvement in Photovoltaic Energy System
(Vysoká škola báňská - Technická univerzita Ostrava, 2025) Maklad, Ahmed M; Morsy, Gamal A.; Khattab, Heba A; Amer, Ragab A.
The growing use of advanced equipment in modern systems, such as electronic devices and drives, has led to a decline in power quality (PQ), causing malfunctions in sensitive loads. Additionally, the integration of renewable energy sources into the power grid significantly impacts the PQ of the electrical system. To address these effects on voltage stability and harmonic distortion, the unified power flow controller (UPFC) series compensator has proven to be a highly effective solution. This study focuses on using the UPFC to mitigate PQ issues related to renewables, including voltage sag, swell, harmonics, and fault conditions. The UPFC is controlled by a fractional order proportional integral derivative (FOPID) controller, which uses the improved zebra optimization algorithm (ZOA) to determine optimal gain values under various PQ scenarios. Furthermore, three comparative assessments of different optimization approaches are conducted to achieve the desired performance and power of the proposed UPFC. The results showed that the proposed ZOA approach compared with WOA and PSO yielded the shortest computing time of 173.554, 257.544, and 382.405 seconds and achieved an objective function value of 2.371, 2.387, and 2.398, respectively. The effectiveness of the proposed strategy is validated using the MATLAB/Simulink platform, with results showing significant improvements in voltage stability and harmonic reduction.
Performance Comparisons between Conventional and Hairpin Winding Configurations of V-Shaped IPMSM
(Vysoká škola báňská - Technická univerzita Ostrava, 2025) Quoc, Vuong Dang; Minh, Dinh Bui; Duc, Hung Bui; Chi, Phi Do
Interior permanent magnet synchronous motors (IPMSMs) have been used extensively in the transportation, industrial, medical, and military fields recently because of their many benefits, including high power density, high torque, and high operational economy. However, current studies have mainly focused on using traditional windings for these motors, neglecting the potential improvements in motor performance that could be achieved with different winding types. Therefore, it is crucial to evaluate the performance of these motors when using different winding structures. This paper presents a combination of analytical technique and finite element method to compare the electromagnetic parameters (back electromotive force (EMF), output power and torque, and temperature rise) of IPMSMs with conventional and hairpin winding configurations. The validated method is then applied to a practical V-shape IPMSM with conventional and hairpin winding configurations.
Smart Military Logistics Based on Internet of Things and Energy Harvesting
(Vysoká škola báňská - Technická univerzita Ostrava, 2025) Malik, Mandee; Kothari, Ashwin; Pandhare, Rashmi
Logistics is a challenge in military as things are fluid and on the move constantly 24 X 7. Managing inventory for weapon systems, vehicle parts, ammunition, food items, utensils, clothing and other day-today necessary items is cumbersome and needs a professional solution incorporating the geographical challenges faced by military. Supply chain in military is done manually and thus prone to errors due to poor inventory management and manual documentation. This paper brings a green solution to this problem with minimum resource requirement and maximum coverage, also keeping in mind the remote area challenges such as power and weather conditions. Backscatter radios communicate by means of reflections without any batteries; rather they generate energy from ambient electromagnetic waves through energy harvesting. Multistatic scatter network uses carrier emitters and readers to gather information from hundreds of tags placed on different set of items spread over a large area. This data generated by reader from tags in a warehouse can automate the inventory management and lead to smart logistics system for military cargo facilities. This paper proposes a multi-static topology for large warehouses with non-linear energy harvesting for effective coverage of cargo facility. Paper also gives an insight on information and energy outage in small-scale fading scenarios. Comparison with linear energy harvesting and mono-static model is also carried out and substantiated by simulations.
Enhancing IEEE 802.11ah Networks: A Spatial Multi-Channel MAC Protocol
(Vysoká škola báňská - Technická univerzita Ostrava, 2025) Dang, Duc Ngoc Minh
IEEE 802.11ah is designed to enhance IoT networks by supporting numerous stations, extending coverage range, reducing power consumption, and operating within the sub-1 GHz band. The Restricted Access Window is introduced to mitigate collisions and improve network throughput when multiple stations contend for the channel simultaneously. While the PHY layer in IEEE 802.11ah supports multiple channels with different bandwidths, the MAC layer only supports a single channel for communication between the Access Point (AP) and stations. However, network performance can be improved by enabling stations at different locations to exchange data packets on different channels concurrently. This paper proposes a Spatial Multi-channel MAC (SM-MAC) protocol for IEEE 802.11ah, which divides the AP’s coverage area into sectors, each assigned a dedicated channel. Each sector is served by two Forwarders responsible for relaying data packets between stations and the AP. The performance of the SM-MAC protocol is evaluated against the IEEE 802.11ah MAC protocol in terms of packet delivery ratio, throughput, and energy efficiency.
Modeling of Asynchronous Units of a Powerful Pumping Station in the Pressure Stabilization Mode
(Vysoká škola báňská - Technická univerzita Ostrava, 2025) Lysiak, Vladyslav
The paper is devoted to developing a mathematical model of a powerful controlled pumping station with automatic pressure stabilization. The work aims to increase the efficiency of the liquid transportation process through main and large distribution pipelines. The proposed model with a comparable level of detail describes the electromechanical and hydraulic subsystems as a single entity. The equations of the hydraulic subsystem are formed based on the principle of electrohydrodynamic analogy and reflect the physical processes in its components. The parameters of the centrifugal pump model were calculated based on the geometry and dimensions of its internal elements, taking into account the influence of the physical properties of the working fluid. It makes it possible to conduct a comprehensive study of such objects without physical impact on them, taking into account the mutual dependence of both subsystems and changes in the parameters of the elements of the hydraulic subsystem. The paper offers directions for use and ways to improve the functionality of the developed model.
Multi-Power Beacon Empowered Secure In Iot Networks: Secrecy Outage Probability Analysis
(Vysoká škola báňská - Technická univerzita Ostrava, 2025) Hung, Tran Cong; Nguyen, Quang Sang; Minh, Bui Vu; Nguyen, Thu-Quyen Thi; Nguyen, Ngoc-Long
This paper investigates the physical-layer secure performance of a multi-power beacon-empowered wirelessly powered communication system in the surveillance of an external eavesdropper. Specifically, a limited-energy source harvests energy from multiple dedicated power beacons in the first time slot and reuses it to perform data transmission in the second time slot. However, its communication with the destination is wiretapped by a potentially idle untrusted user. Thus, to evaluate the secure performance of the considered system, we have derived closed-form expressions for the secrecy outage probability (SOP) metric in terms of both exact and asymptotic aspects. Through these formulas, we then provide a series of numerical discussions on how to configure the number setting of the power beacon, the energy harvesting parameters as well as scheduling time to improve the SOP performance, where the Monte-Carlo simulation method is used to verify the developed mathematical framework.
How Generative AI and the Intelligent Industrial Internet of Things Complement Each other
(Vysoká škola báňská - Technická univerzita Ostrava, 2025) Sikora, Axel
Generative modeling is an artificial intelligence (AI) technique to generate synthetic artifacts from analyzing training examples and from learning their patterns and distribution. Generative AI (GenAI) uses generative modeling and advances in Deep Learning to produce diverse content at scale by utilizing existing data. Whereas traditionally GenAI is mostly using media contents, such as text, graphics, audio, and video, it can additionally be used also for data from the (Industrial) Internet of Things. This article provides a systematic overview on the manifold different practical opportunities and challenges GenAI brings for the IIoT. It also presents selected examples from the author’s research with his teams. In doing so, it covers the relevance of GenAI for the complete lifecycle of IIoT, from design and development, over testing to deployment. This paper summarizes a keynote presentation from the 13th International Conference on Green and Human Information Technology (ICGHIT) in January 2025 held in Nha Trang, Vietnam.
The Impacts of Electric Charging Stations on Distribution Power Grids under Different Simulations using Jellyfish Swarm Algorithm
(Vysoká škola báňská - Technická univerzita Ostrava, 2025) Tran, Dao Trong; Duong, Minh Phuc
This research presents different implementations for placing electric charging stations (ECSs) in distribution power networks (DPNs) to achieve the best total active power loss (TAPL). Solar generators (SGs) are also used to alleviate the adverse effects resulting from the presence of ECSs in the networks in terms of power loss and voltage profile. Artificial hummingbird algorithm (AHA), Jellyfish swarm algorithms (JS), and Northern goshawk optimization (NGO) are executed to determine the best placement of ECSs and SGs in an IEEE 33-node network for reaching a minimum loss and satisfying all the related constraints. There are four cases conducted in the whole research. In the first case, JS outperforms both AHA and NGO by providing the highest stability throughout all the trial runs and fastest convergence speed to the optimal solution in the best runs. Besides, the quantitative comparison also consolidates the robustness and reliability of JS compared to others. Based on the surprising performance, JS is continuously reapplied to solve another three cases of the considered problem. Through those three cases with the application of JS, the TAPL values of four scenarios with different numbers of ECSs are evaluated. Specifically, the results achieved by JS indicate that the higher number of ECSs leads to a higher value of TAPL and a higher voltage drop. On the other hand, the simultaneous placement of SGs and ECSs can result in smaller fluctuations of the voltage profile and smaller TAPL. Thus, the optimization placement of ECSs and SGs is crucial to DPNs for economic and technical purposes.

Browse

Recent Submissions