Original Research Paper
Artificial Intelligence
Hossein KardanMoghaddam; Adel Akbarimajd; Shahram Jamali; Hamid KardanMoghaddam
Abstract
Background and Objectives: Multi-agent systems that incorporate large language models (LLM-MAS) represent one of the most innovative areas in artificial intelligence. These systems effectively address significant problems across various fields by merging the capabilities of large language models with ...
Read More
Background and Objectives: Multi-agent systems that incorporate large language models (LLM-MAS) represent one of the most innovative areas in artificial intelligence. These systems effectively address significant problems across various fields by merging the capabilities of large language models with traditional multi-agent systems (MAS). As science progresses in various domains, the importance of natural language processing, particularly text classification has significantly increased.Methods: This study introduces an innovative framework for text classification based on a multi-agent voting system. In this framework, four distinct LLMs independently label input text into five predefined categories, assigning a weight to each label. Additionally, five specialized term extraction functions act as independent agents, identifying key concepts related to each label within the text. These functions influence the final decision-making process by adjusting the weights assigned by the language models. The label with the highest aggregated weight is selected as the final classification output.Results: Experimental results demonstrate that the proposed framework achieves an accuracy of approximately 86%, highlighting the effectiveness of an LLM-based multi-agent approach in text classification tasks.Conclusion: Fine-tuning large language models typically requires a substantial dataset to effectively personalize the model. This process can be time-consuming and costly, and the resulting model is often limited to a specific application. However, the results of this study demonstrate that the proposed framework can achieve acceptable accuracy without the need for fine-tuning large language models. The study shows that combining term extraction functions with large language models can create smarter and more accurate systems. These term extraction functions modify the weights derived from the large language models, adjusting them toward specific topic labels based on relevant words found in the text.
Original Research Paper
Modelling, simulation and verification
Hossein Rashidi; Hossein Khaleghi
Abstract
Background and Objectives: Massive Multiple-Input Multiple-Output (MIMO) systems operating in the millimeter-wave (mmWave) frequency bands offer high data rates and spectral efficiency, but accurate channel modeling remains challenging due to complex propagation characteristics. Existing channel models ...
Read More
Background and Objectives: Massive Multiple-Input Multiple-Output (MIMO) systems operating in the millimeter-wave (mmWave) frequency bands offer high data rates and spectral efficiency, but accurate channel modeling remains challenging due to complex propagation characteristics. Existing channel models vary widely, and a unified framework for efficient channel estimation in next-generation systems is still lacking. This study aims to develop a simplified yet accurate channel model and a corresponding channel estimation method tailored for massive MIMO mmWave systems.Methods: A novel channel model is proposed based on the Saleh–Valenzuela (S-V) model and formulated using the Discrete Fourier Transform (DFT) to capture the spatial characteristics of mmWave channels efficiently. Leveraging this model, a DFT-based channel estimation method is developed, designed to reduce computational complexity while maintaining high accuracy. The performance of the proposed approach is evaluated through numerical simulations under various large-scale antenna array configurations.Results: Simulation results demonstrate that the proposed DFT-based channel model accurately approximates mmWave propagation characteristics and achieves higher channel estimation accuracy than conventional methods. Additionally, the method significantly reduces computational complexity and processing time. As the number of antennas increases, the performance of the proposed approach converges closely to that of traditional S-V models, confirming its scalability and effectiveness for massive MIMO scenarios.Conclusion: The study presents a practical and efficient framework for channel modeling and estimation in massive MIMO mmWave systems. The proposed DFT-based model and estimation method provide a balance between accuracy and computational efficiency, offering a valuable tool for the design and optimization of next-generation wireless networks.
Original Research Paper
Micro Sensors
Morteza Janfaza; Hamed Moradi
Abstract
Background and Objectives: The domain of optical fiber acoustic sensors (OFAS) has witnessed considerable progress, with a specific emphasis on enhancing their sensitivity. Among the approaches explored, optical phase shift detection and the implementation of tapered single-mode fibers (SMFs) have garnered ...
Read More
Background and Objectives: The domain of optical fiber acoustic sensors (OFAS) has witnessed considerable progress, with a specific emphasis on enhancing their sensitivity. Among the approaches explored, optical phase shift detection and the implementation of tapered single-mode fibers (SMFs) have garnered particular attention. This study evaluates the sensitivity characteristics of interferometric fiber optic acoustic sensors possessing bare SMF, tapered SMF, and standard SMF, all configured within a Michelson interferometer. Methods: To achieve an efficient hardware implementation, the arctangent is calculated using the Coordinate Rotation Digital Computer (CORDIC) algorithm. CORDIC algorithm computes trigonometric functions through an iterative process of vector rotation. Its principal advantage lies in its reliance on elementary operations—specifically, additions, bit shifts, and table look-ups—which renders it a divider-free and highly area-efficient solution for nonlinear function evaluation.Results: This work employs a Michelson interferometer configuration, which is illuminated by a 10 mW, 1550 nm Distributed Feedback Laser Diode (DFB-LD). The sensor probe integrates three distinct 1 cm lengths of standard, tapered (30 μm diameter), and bare single-mode fiber. This fiber assembly is embedded in a low-refractive-index polymer gel medium and mounted rigidly onto a metal plate. A piezoelectric phase modulator in the reference arm is driven by a 33 kHz signal, while acoustic stimuli are generated by a speaker, enabling a comparative performance analysis with a co-located commercial microphone. The findings indicate that bare SMF offers enhanced detectability for acoustic disturbances relative to its tapered and standard counterparts. The sensor is capable of resolving acoustic signals across 500 Hz to 4500 Hz.Conclusion: The demodulated output of an OFAS is presented, confirming the system's ability to isolate acoustic signals from environmental drift. The acoustic sensitivity of three fiber configurations—standard SMF, tapered SMF (30 µm diameter), and bare SMF—was experimentally investigated. Findings demonstrate that the bare SMF exhibits significantly greater sensitivity than its tapered and standard counterparts. This enhanced response was corroborated by a strong correlation between the fiber sensor's output and the signal from a reference standard microphone.
Original Research Paper
Bioinformatics
Mahboubeh Ayoubi; Babak Teimourpour; Mostafa Akhavan-Safar
Abstract
Background and Objectives: Identifying and classifying cancer-driving genes by analyzing their complex relationships within gene regulatory networks (GRN) can significantly aid in the development, progression, and discovery of targeted cancer therapies. The cancer-driving genes are responsible for tumorigenesis ...
Read More
Background and Objectives: Identifying and classifying cancer-driving genes by analyzing their complex relationships within gene regulatory networks (GRN) can significantly aid in the development, progression, and discovery of targeted cancer therapies. The cancer-driving genes are responsible for tumorigenesis and disease progression. However, the current methods frequently concentrate on network rebuilding, which restricts their capacity to identify regulatory linkages. By utilizing the structural and functional characteristics of GRNs, this work seeks to create a strong graph-based framework for precise cancer driver gene classification.Methods: Network and graph-based methodologies are employed to analyze these complex gene networks. Using graph neural networks (GNN), complex intergenic patterns can be identified in genetic and cellular data. In this study, a GNN-based framework is proposed to classify genes in gene regulatory networks in order to improve the detection of cancer-driving genes. The proposed graph-based framework effectively integrates multi-omics data and mitigates class imbalance through an artificial oversampling strategy. The proposed GNN-based framework facilitates the modeling of both topological structure and feature information within gene interaction networks. Additionally, it addresses the challenges of class imbalance between driver and non-driver genes through the implementation of the GraphSMOTE technique.Results: To construct the gene regulatory graph, the Regetworks regulatory dataset was combined with three gene expression datasets related to breast, lung, and colon cancers. The results demonstrate that the proposed model consistently attains robust classification performance, with AUC-ROC scores exceeding 0.77 in all cases and F1 scores above 0.70, outperforming previous network-based methods.Conclusion: The evaluation criteria show that the proposed model has a high ability to generalize tumor types with differences in network topology and class imbalance.
Original Research Paper
Cloud Computing
Gowri S; Jaganathan Rathi
Abstract
Background and Objectives: Cloud computing can play a vital role in promoting environmental sustainability by leveraging eco-friendly dedicated servers that adhere to green computing standards. The concept of "green cloud computing" revolves around harnessing cutting-edge technologies to minimize the ...
Read More
Background and Objectives: Cloud computing can play a vital role in promoting environmental sustainability by leveraging eco-friendly dedicated servers that adhere to green computing standards. The concept of "green cloud computing" revolves around harnessing cutting-edge technologies to minimize the environmental footprint of computing systems. One of the significant challenges in cloud-based systems is task scheduling, which must be optimized to enhance system efficiency, user experience, and environmental sustainability.Method: This paper proposes a novel Hybrid HEES (Hierarchical Energy-Efficient Scheduling) method that optimizes energy consumption and task scheduling in cloud computing environments. By combining genetic algorithm optimization, workflow-based scheduling, and energy-aware resource allocation, HEES achieves significant reductions in energy consumption and average task completion time.Results: The method is evaluated through simulations, demonstrating its effectiveness in optimizing energy efficiency and task scheduling performance. The Hybrid HEES method has the potential to reduce energy consumption, improve computing performance, and enhance sustainability in cloud computing environments.Conclusion: To evaluate a proposed HEES method through cloudsim 3.0 simulations, the numerical results confirm the effectiveness of HEES algorithm, which achieves average Energy consumption performance improvements of around 12% compared to GP and 8% compared to RR existing methods.
Original Research Paper
Power Systems
Ramazan Havangi; Fatemeh Karimi
Abstract
Background and Objectives: Accurate state-of-charge (SOC) estimation is essential for improving the performance, reliability, and lifetime of lithium-ion battery systems. Although neural-network-based methods have demonstrated promising SOC estimation capability, the influence of excitation signal frequency ...
Read More
Background and Objectives: Accurate state-of-charge (SOC) estimation is essential for improving the performance, reliability, and lifetime of lithium-ion battery systems. Although neural-network-based methods have demonstrated promising SOC estimation capability, the influence of excitation signal frequency characteristics on estimation performance has not been systematically investigated. This study aims to address this gap by examining how the frequency content of excitation signals influences the SOC estimation performance of NNs.Methods: An Amplitude Pseudo-Random Binary Sequence (APRBS) signal was applied as the excitation input to the battery system. Fast Fourier Transform (FFT) analysis was then conducted to assess the frequency content of the APRBS signal. The relationship between the frequency components of the APRBS input and the SOC estimation error of the NN was systematically investigated.Results: The results indicate that increasing the frequency bandwidth of the APRBS input signal significantly improves the SOC estimation accuracy of the neural network, particularly during the initial stages of operation. However, beyond a certain bandwidth threshold, the improvement becomes marginal, indicating the existence of an optimal frequency range that provides the best trade-off between excitation richness and estimation performance. Furthermore, a correlation was identified between the extracted frequency components and the Bode diagram of the battery system, providing valuable insights into the underlying dynamic characteristics of the battery.Conclusion: The findings demonstrate that the frequency characteristics of input signals play a critical role in SOC estimation accuracy when using NNs. Identifying an optimal frequency bandwidth not only improves estimation performance but also enhances understanding of battery dynamics. This work introduces a novel perspective for optimizing SOC estimation through the integration of signal processing and machine learning, laying the groundwork for future advancements in battery management systems.
Original Research Paper
Artificial Intelligence
Zohre Moteshakker Arani; Mahdi Naghibi
Abstract
Background and Objectives: Modern LLMs struggle with retaining long-term knowledge, personalized context, and complex reasoning over extended texts. A surge of memory-augmented LLM architectures addresses this by integrating external memory mechanisms. Broadly, these methods include retrieval-augmented ...
Read More
Background and Objectives: Modern LLMs struggle with retaining long-term knowledge, personalized context, and complex reasoning over extended texts. A surge of memory-augmented LLM architectures addresses this by integrating external memory mechanisms. Broadly, these methods include retrieval-augmented generation (RAG) (augmenting LLMs with external documents or knowledge bases), explicit memory modules (learned read/write stores), long-context transformers (architectural extensions for longer inputs), episodic/personal memory (storing user or session history), and neuro-symbolic memory (knowledge-graph or symbolic integration). Methods: In this survey, we present a technical overview of how short-term, long-term, and other memory mechanisms are implemented in and around LLMs (including multimodal agents), and how these innovations enable more capable AI systems. We survey key journal and peer-reviewed works (2021–2025) in each category, emphasizing advances in reasoning and assistant-style tasks. We compare prominent LLM systems that incorporate memory, discuss applications in agents and continuous dialogues, explore the use of memory in different applications, and highlight challenges and future research directions in memory-augmented LLMs.Results: We contribute a comparative mapping between memory types and representative application domains—such as conversational assistants, knowledge-intensive reasoning, scientific research support, and personalized systems—to address the central question of which memory strategy best aligns with which task requirements. Conclusion: Although considerable advances have been achieved, important challenges remain. The literature suggests that increasingly sophisticated memory mechanisms will play a central role in the development of next-generation LLMs and AI agents, necessitating future research directions toward more robust, adaptive, and scalable memory-augmented language model systems.
Original Research Paper
Electronic Circuits
Bahram Rashidi; Leila Golpaiegany
Abstract
Background and Objectives: This research presents a circuit design for controlling the speed and direction of high-current Direct Current (DC) motors. The circuit can control DC motors with voltages ranging from 12V to 48V and currents up to 60A. Given that the switching MOSFETs in 48V motors operate ...
Read More
Background and Objectives: This research presents a circuit design for controlling the speed and direction of high-current Direct Current (DC) motors. The circuit can control DC motors with voltages ranging from 12V to 48V and currents up to 60A. Given that the switching MOSFETs in 48V motors operate at high voltages, the circuit must be capable of handling these conditions. Various techniques are employed in the design of the proposed circuit to control these high-voltage motors effectively.Methods: The proposed hardware implementation incorporates several well-established techniques in power electronics to enhance reliability, switching performance, and protection under high-current operating conditions. These include: (1) ultra-fast parallel diodes for motor freewheeling current paths, (2) RC snubber networks across each MOSFET to suppress voltage overshoot and oscillations during switching, (3) a series gate resistor combined with a reverse diode to ensure controlled gate charging and provide a fast discharge path, (4) Schottky diodes connected across the MOSFET terminals to improve switching behavior and reduce stress, (5) an opto-isolator to provide electrical isolation between the control and power stages, and (6) the use of four parallel MOSFETs in the switching stage to improve current handling capability and reduce conduction losses. To accommodate different motor characteristics, the output pulse frequency can be adjusted using a variable capacitor. In addition, smooth startup operation is achieved by initially setting the PWM duty cycle to a low value and gradually increasing it, thereby preventing abrupt motor acceleration and reducing mechanical shock during startup.Results: The pulse width modulation (PWM) pulse frequency generated by the pulse generator is 6.356 kHz, with a duty cycle that can be adjusted from 3% to 99%. Additionally, we have implemented a variable-frequency mode, enhancing the circuit's ability to control a range of motors.Conclusion: Based on the tests conducted, the proposed circuit can effectively control the speed and direction of high-current DC motors without issues. Considering the components used, the circuit is capable of supplying the necessary power to drive DC motors operating at 12-48 V and currents up to 60 A.
Original Research Paper
Artificial Intelligence
Seyed Hamid Zahiri; Mahdi Moodi
Abstract
Background and Objectives: An Android botnet refers to a collection of infected Android devices under the control of a cybercriminal, known as a botmaster, which are used to carry out harmful actions like DDoS attacks, spam distribution, data theft, or malware propagation. These botnets present significant ...
Read More
Background and Objectives: An Android botnet refers to a collection of infected Android devices under the control of a cybercriminal, known as a botmaster, which are used to carry out harmful actions like DDoS attacks, spam distribution, data theft, or malware propagation. These botnets present significant risks to user security, privacy, and network stability. Detecting them is difficult due to their adaptive nature, as attackers continually refine their methods to avoid detection. Additionally, Android devices have resource limitations—such as battery life, memory, and processing capacity—that further complicate detection efforts. Methods: To address this challenge, this paper proposes the SGO DF (Squid Game Optimization-based Data Fusion) method for accurately identifying Android botnets. The approach optimizes eleven parameters using the Squid Game Optimization (SGO) algorithm: eight of these parameters fine-tune machine learning models (SVM, DT, and RF) to enhance their accuracy, while the remaining three adjust weighting in the Data Fusion process. Results: This optimization enables more precise classification of test data, outperforming five other methods in evaluation. Conclusion: The results demonstrate that SGO DF achieves an average accuracy exceeding 99% across 28 different Android botnet datasets, showcasing its effectiveness in botnet detection.
