<?xml version="1.0" encoding="utf-8"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom">
    <channel>
        <title>Journal of Electrical Engineering Feed</title>
        <link>https://reference-global.com/journal/JEE</link>
        <description>Reference Global RSS Feed for Journal of Electrical Engineering</description>
        <lastBuildDate>Thu, 22 Oct 2026 20:09:19 GMT</lastBuildDate>
        <docs>https://validator.w3.org/feed/docs/rss2.html</docs>
        <generator>https://github.com/jpmonette/feed</generator>
        <image>
            <title>Journal of Electrical Engineering Feed</title>
            <url>https://paradigm-files.s3.eu-central-1.amazonaws.com/JEE/cover-image.jpg</url>
            <link>https://reference-global.com/journal/JEE</link>
        </image>
        <copyright>All rights reserved 2026, Slovak University of Technology in Bratislava</copyright>
        <atom:link href="https://reference-global.com/2/rss/journal/JEE" rel="self" type="application/rss+xml"/>
        <item>
            <title><![CDATA[CFOA-based mixed-mode first-order universal filter implemented with commercially available ICs]]></title>
            <link>https://reference-global.com/article/10.2478/jee-2026-0047</link>
            <guid isPermaLink="false">/article/10.2478/jee-2026-0047</guid>
            <pubDate>Sun, 11 Oct 2026 00:00:00 GMT</pubDate>
            <description><![CDATA[This paper introduces a mixed-mode first-order universal filter, which is implemented using commercially available integrated circuits. The proposed filter comprises three current feedback operational amplifiers (CFOAs) as commercially available ICs, along with three resistors and one capacitor as passive components. The three standard responses of a first-order filter, i.e., lowpass, highpass, and allpass responses, can be accomplished simultaneously with a single proposed structure. Additionally, it can operate in all four modes of operation, including voltage mode, transadmittance mode, current mode, and transimpedance mode. The circuit enables pole-frequency tunability through a single resistor. A non-ideal analysis considering the CFOA tracking errors and parasitic elements was also investigated. A simple application of the proposed filter for implementing a quadrature oscillator is suggested. To confirm the performance of the filter, the simulation results using PSPICE software were performed.]]></description>
            <category>ARTICLE</category>
        </item>
        <item>
            <title><![CDATA[CMOS rail-to-rail voltage-to-current converter circuit intended for low-voltage IC applications]]></title>
            <link>https://reference-global.com/article/10.2478/jee-2026-0046</link>
            <guid isPermaLink="false">/article/10.2478/jee-2026-0046</guid>
            <pubDate>Sun, 11 Oct 2026 00:00:00 GMT</pubDate>
            <description><![CDATA[In this paper a new CMOS voltage-to-current converter (VTC) circuit with rail-to-rail dynamic range is proposed. The circuit is dedicated for integrated circuits that require low-voltage power supply. The input/output voltage dynamic range of conventional VTC circuits is usually impaired by the reduced of power supply voltage. The proposed VTC aims to alleviate non-desirable the effects of the reduction of the power supply voltage on the VTC dynamic range by using a combination of voltage followers. It is designed using the transistor models of the 65 nm CMOS technology and it is intended to operate with low power supply voltages. SPICE simulation results demonstrate that the input dynamic range of the proposed VTC circuit covers the entire range from 0 V to 1.2 V, the maximum range allowed by the used power supply. This rail-to-rail property makes the proposed VTC circuit highly suitable for CMOS integrated circuits low-voltage applications.]]></description>
            <category>ARTICLE</category>
        </item>
        <item>
            <title><![CDATA[Characterization of Bragg spectrum shifts in cascaded FBG arrays for thermal sensing]]></title>
            <link>https://reference-global.com/article/10.2478/jee-2026-0051</link>
            <guid isPermaLink="false">/article/10.2478/jee-2026-0051</guid>
            <pubDate>Sun, 11 Oct 2026 00:00:00 GMT</pubDate>
            <description><![CDATA[This paper investigates the temperature-sensing performance of a series-configured Fiber Bragg Grating (FBG) sensing network integrated within an optical communication system. Two homogeneous FBG sensors were modelled using OptiSystem 22.0 and monitored using a Wavelength Division Multiplexing (WDM) interrogator. Linear regression analysis yielded a temperature sensitivity of 1.67 pm/°C (R2=0.999), demonstrating an excellent linear relationship between temperature and Bragg wavelength shift. The results demonstrate a positive relationship between temperature and wavelength shift, indicating the capability of cascaded FBG sensors for multipoint thermal monitoring applications. The findings highlight the potential of series-connected FBG arrays for integration into future optical sensing and communication networks.]]></description>
            <category>ARTICLE</category>
        </item>
        <item>
            <title><![CDATA[Secure GNSS telemetry on ESP32 using kinematic validation, adaptive logging, and hardware-accelerated AES encryption]]></title>
            <link>https://reference-global.com/article/10.2478/jee-2026-0052</link>
            <guid isPermaLink="false">/article/10.2478/jee-2026-0052</guid>
            <pubDate>Sun, 11 Oct 2026 00:00:00 GMT</pubDate>
            <description><![CDATA[Global navigation satellite system (GNSS) telemetry is vulnerable to spoofing attacks, while conventional tracking systems often store potentially compromised coordinates before performing server-side validation. This paper presents an ESP32-based secure GNSS telemetry system that performs edge-level kinematic validation prior to data encryption and logging. The proposed method combines a reachability constraint for detecting physically impossible coordinate changes, alpha-beta filtering for trajectory smoothing, zero-velocity anchoring for suppressing navigation drift during stops, and adaptive spatial, angular, and time-based sampling. Validated telemetry records are encrypted in real time using hardware-accelerated AES-256 in counter mode, with session-specific initialization vectors and cryptographic key isolation in the ESP32 eFuse memory. A desktop audit application provides local decryption, trajectory reconstruction, distance calculation, and retrospective analysis of the collected data. Software-in-the-Loop experiments achieved 100% recall for injected coordinate jumps exceeding the calculated reachability limit, with a false-positive rate of 0.33%. In the 42 km route test, adaptive logging reduced the log-file size by 68% compared with fixed-rate one-second logging. The hardware-assisted encryption time for one telemetry record did not exceed 45 µs, which confirms the suitability of the proposed architecture for real-time embedded applications. The developed solution provides a resource-efficient approach to improving the plausibility and confidentiality of GNSS telemetry while supporting retrospective analysis in fleet monitoring and other autonomous tracking systems.]]></description>
            <category>ARTICLE</category>
        </item>
        <item>
            <title><![CDATA[Diagnostic-aware polar coding and adaptive hybrid automatic repeat request with reinforcement learning for hemodynamic telemonitoring of PTSD in 5G/IoT networks]]></title>
            <link>https://reference-global.com/article/10.2478/jee-2026-0053</link>
            <guid isPermaLink="false">/article/10.2478/jee-2026-0053</guid>
            <pubDate>Sun, 11 Oct 2026 00:00:00 GMT</pubDate>
            <description><![CDATA[Wearable telemonitoring of post-traumatic stress disorder (PTSD) relies on heart rate variability derived from the sphygmogram, where the diagnostic conclusion rests on the temporal positions of a few fiducial points, a distinction to which conventional channel metrics are blind. The paper introduces a diagnostic weight of a sparse representation coefficient, obtained from the sensitivity of the fiducial points to a perturbation of that coefficient, and shows that the energy ordering and the diagnostic ordering of the discrete cosine transform (DCT) are opposite. A diagnostic integrity criterion, measured as the deviation of the SDNN index, replaces the bit error rate (BER) as the optimization target. The reliability gradation that a polar code creates among its information subchannels carries the most diagnostically significant bits on the most reliable subchannels, which yields unequal error protection (UEP) at no cost in redundancy or code rate. The retransmission decision follows an estimate of diagnostic risk from soft decoder outputs, with the threshold adapted by a reinforcement learning (RL) agent. Simulation over a multipath non-line-of-sight (NLOS) channel gives a margin of 1.5 dB at equal diagnostic fidelity, a near-threefold reduction of dicrotic notch detection failures at equal block error rate (BLER), and a quarter fewer retransmissions.]]></description>
            <category>ARTICLE</category>
        </item>
        <item>
            <title><![CDATA[Hardware implementation of AI-based fuzzy logic for real-time sign language interpreter robotic control using FPGA]]></title>
            <link>https://reference-global.com/article/10.2478/jee-2026-0049</link>
            <guid isPermaLink="false">/article/10.2478/jee-2026-0049</guid>
            <pubDate>Sun, 11 Oct 2026 00:00:00 GMT</pubDate>
            <description><![CDATA[Communication barriers between hearing-impaired individuals and non-sign language users remain a significant challenge in modern society. Existing sign language recognition systems primarily focus on unidirectional communication and are often limited by high latency, excessive power consumption, and lack of robustness in uncertain environments. This paper proposes a novel FPGA-based Bidirectional Robotic Sign Language Communication System (BRSLCS) integrating Artificial Intelligence (AI), fuzzy logic, and robotic control for real-time assistive communication. The proposed architecture supports both forward communication, where sign language gestures are translated into text or speech, and reverse communication, where text or speech is converted into robotic sign language gestures through an actuated robotic hand. In the forward communication flow, gesture images captured using a camera are processed through preprocessing, feature extraction, AI-based classification, and fuzzy inference stages. In the reverse communication flow, speech or textual inputs are interpreted using Natural Language Processing (NLP) techniques and mapped into robotic sign gestures controlled using FPGA-based actuation logic. The integration of fuzzy logic improves classification robustness and interpretability under uncertain conditions such as illumination variation, gesture ambiguity, and motion blur. FPGA implementation provides deterministic timing, low latency, hardware parallelism, and energy efficiency suitable for embedded real-time systems. Experimental evaluation demonstrates that the proposed system achieves 92-95% recognition accuracy with latency between 10-25 ms while consuming only 4-6 W power, outperforming conventional CPU and GPU-based implementations. The proposed BRSLCS provides a scalable and practical framework for inclusive communication systems in healthcare, education, public services, and human-robot interaction environments.]]></description>
            <category>ARTICLE</category>
        </item>
        <item>
            <title><![CDATA[A neural net fitting-based approach for optimal power flow]]></title>
            <link>https://reference-global.com/article/10.2478/jee-2026-0055</link>
            <guid isPermaLink="false">/article/10.2478/jee-2026-0055</guid>
            <pubDate>Sun, 11 Oct 2026 00:00:00 GMT</pubDate>
            <description><![CDATA[Optimal power flow (OPF) is one of the most important management actions to achieve economical, safe and reliable operation of power systems. Require both high accuracy and speed of calculation so that appropriate corrective control actions can be implemented in real time. Classical methods such as interior-point method, sequential programming methods, as well as metaheuristic optimization methods used today require a lot of time to achieve satisfactory accuracy of OPF results. With the development of artificial intelligence and neural networks, it is possible to drastically reduce the OPF calculation time. This paper proposes the application of Neural Net Fitting (NNF) based approach for OPF. The basic idea of this work is to use a combination of metaheuristic optimization methods (MOA) and the NNF technique to realize a kind of hybrid method-network, which would solve the OPF problem in a very efficient and fast way, in a way suitable for real-time application. The proposed approach is tested on the standard IEEE 30-bus test system. The simulation results show that the application of the proposed NNF technique yields the same accuracy OPF results as the application of the classical metaheuristic methods PSOGSA or GA with the hybrid fmincon interior-point function in MATLAB®, but with significantly less computation time.]]></description>
            <category>ARTICLE</category>
        </item>
        <item>
            <title><![CDATA[Proof of concept for DNA storage system with noise-guessing decoder]]></title>
            <link>https://reference-global.com/article/10.2478/jee-2026-0050</link>
            <guid isPermaLink="false">/article/10.2478/jee-2026-0050</guid>
            <pubDate>Sun, 11 Oct 2026 00:00:00 GMT</pubDate>
            <description><![CDATA[DNA storage has significant advantages if compared with existing electronic storage systems, usually concentrated in data centers, which form digital cloud infrastructure. These are: longevity, format used by nature, high capacity per volume and small CO2 footprint. Information which is stored and later retrieved from DNA storage systems can be corrupted by insertions, deletions and substitutions (INDELS). Oxford nanopore sequencing is probably the closest to making DNA storage a mass product. On the other hand, different decoding systems based on noise-guessing paradigm also possess advantages when compared with classical error control coding, which has been demonstrated for wide variety of communications channels. Therefore, in this manuscript a proof of concept is given for a DNA storage system with a decoder inspired by the noise-guessing paradigm approach. It is given using a simple model for INDELS obtained from real measured data.]]></description>
            <category>ARTICLE</category>
        </item>
        <item>
            <title><![CDATA[Energy-efficient coverage and connectivity maintenance in UAV swarm networks via multi-agent coordination]]></title>
            <link>https://reference-global.com/article/10.2478/jee-2026-0048</link>
            <guid isPermaLink="false">/article/10.2478/jee-2026-0048</guid>
            <pubDate>Sun, 11 Oct 2026 00:00:00 GMT</pubDate>
            <description><![CDATA[Time-constrained disaster management with micro-unmanned aerial vehicle (UAV) swarms requires autonomous coordination to ensure effective spatial coverage while preserving reliable ad-hoc network connectivity. The inherent trade-off between coverage and connectivity is compounded by the strict aerodynamic power limitations of micro-UAVs, whose batteries deplete rapidly under erratic maneuvers. Conventional centralized control architectures and standard artificial potential field (APF) algorithms often fail to address this multi-objective problem, leading to fragmented networks or wasted energy due to oscillatory trajectories (jitter). To overcome these limitations, this paper proposes the Energy-Damped Virtual Force (EDVF) algorithm, a fully decentralized, neighbor-aware control law. EDVF employs an adaptive velocity-dependent damping term (analogous to viscous friction) that penalizes unnecessary micro-corrections as drones settle into their topological positions. We formulate a multi-objective optimization model that integrates algebraic connectivity with a non-linear aerodynamic power model based on Blade-Element Momentum Theory. Python-based simulations demonstrate that EDVF outperforms both traditional artificial potential field methods and centralized benchmarks: it maintains approximately 39% percent area coverage – 99% percent of the theoretical maximum for the swarm configuration – guarantees a connected network throughout the mission λ2 &gt; 0, and demonstrably extends swarm operational lifetime by eliminating kinematic oscillations and driving agent velocities asymptotically to zero, thereby removing the cubic aerodynamic drag penalty P_par ∝ ||v||³ that undamped baselines continuously incur. EDVF thus provides a theoretically grounded and scalable solution for maintaining persistent aerial surveillance in stochastic environments.]]></description>
            <category>ARTICLE</category>
        </item>
        <item>
            <title><![CDATA[Temperature dependent fractional order selfcorrecting model with hysteresis for LFP batteries]]></title>
            <link>https://reference-global.com/article/10.2478/jee-2026-0054</link>
            <guid isPermaLink="false">/article/10.2478/jee-2026-0054</guid>
            <pubDate>Sun, 11 Oct 2026 00:00:00 GMT</pubDate>
            <description><![CDATA[Lithium iron phosphate (LFP) cells are attractive due to their thermal stability, safety, and cycle-life characteristics. However, estimating the state-of-charge (SOC) for LFP cells is challenging because the open-circuit voltage (OCV) curve is relatively flat over a wide middle SOC region. This paper develops and evaluates a fractional-order extension of the enhanced self-correcting (ESC) equivalent-circuit model for LFP cell modelling and SOC estimation. The conventional ESC model represents polarisation using an integer-order (IO) resistor-capacitor branch, which summarises the dynamic voltage response using a finite exponential state. To introduce memory-dependent polarisation dynamics while retaining the ESC structure, the capacitor in the polarisation branch is replaced by a fractional-order element. The resulting fractional-order ESC (FO-ESC) model is evaluated using measured ATL and A123 LFP datasets over eight temperatures. The tuned FO-ESC model produced modest average terminal-voltage RMSE reductions relative to the tuned IO-ESC model. SOC-estimation tests using coulomb counting, OCV lookup, EKF, and SPKF/UKF showed that the FO plant did not destabilise the estimators. The results support FO-ESC as a temperature-dependent refinement of the ESC framework and motivate the development of future FO-aware SOC estimators.]]></description>
            <category>ARTICLE</category>
        </item>
        <item>
            <title><![CDATA[Dynamics-aware local trajectory control of a tracked vineyard robot via deep reinforcement learning]]></title>
            <link>https://reference-global.com/article/10.2478/jee-2026-0045</link>
            <guid isPermaLink="false">/article/10.2478/jee-2026-0045</guid>
            <pubDate>Thu, 27 Aug 2026 00:00:00 GMT</pubDate>
            <description><![CDATA[Autonomous navigation between narrow vineyard rows requires respecting the platform’s dynamics, which trajectory optimizers such as CHOMP and STOMP resolve in batch before motion and again whenever the map changes. We instead cast local navigation as a continuous-control Markov decision process whose actions are the per-track torques of a skid-steer robot, so the platform’s dynamics enter the control law itself, and the optimization is paid once, during training. Soft Actor-Critic (SAC) and Twin Delayed Deep Deterministic policy gradient (TD3), both recurrent, and a feed-forward Proximal Policy Optimization (PPO) baseline are trained over ten seeds in simulation from a real vineyard passability map. On two held-out scenarios all three produce shorter routes than a conservative weighted A* reference at higher peak but comparable average impassability, and the safety ranking of the off-policy agents reverses between scenarios.]]></description>
            <category>ARTICLE</category>
        </item>
        <item>
            <title><![CDATA[Control of nonlinear continuous-time multi-input and multi-output dynamic systems using neuroevolution]]></title>
            <link>https://reference-global.com/article/10.2478/jee-2026-0035</link>
            <guid isPermaLink="false">/article/10.2478/jee-2026-0035</guid>
            <pubDate>Thu, 27 Aug 2026 00:00:00 GMT</pubDate>
            <description><![CDATA[We propose a neural network-based controller of nonlinear multi-input and multi-output (MIMO) continuous-time dynamic systems. The controller learns using a genetic algorithm (GA) to minimize the chosen criterion function. A new neural network (NN) architecture is proposed that uses difference- and summation-based neurons in the hidden layers. Two control approaches are compared: a decentralized approach, which assumes that individual subsystems are isolated feedback loops, and a centralized one, which assumes that the entire system is controlled as a single system including all internal interactions between subsystems. Both approaches are compared experimentally with a conventional decentralized PID controller approach. We will show that both NN approaches can learn the plant non-linearity, however, the centralized control approach can much better eliminate the mutual internal interactions among the partial subsystems of the MIMO system.]]></description>
            <category>ARTICLE</category>
        </item>
        <item>
            <title><![CDATA[HSARA: Heuristic Social-Aware Resource Allocation for robust D2D multicasting in UAV-assisted dense 5G networks]]></title>
            <link>https://reference-global.com/article/10.2478/jee-2026-0041</link>
            <guid isPermaLink="false">/article/10.2478/jee-2026-0041</guid>
            <pubDate>Thu, 27 Aug 2026 00:00:00 GMT</pubDate>
            <description><![CDATA[The rapid proliferation of smart devices and multimedia-intensive applications poses unprecedented demands on 5G mobile infrastructure, especially when the terrestrial base stations (BS) are unavailable or overwhelmed. When used as an aerial BS, Unmanned Aerial Vehicles (UAVs) offer a compelling and flexible way to restore and improve wireless coverage. In this work, we examine the allocation of social-aware resources for multicast device-to-device (D2D) communications under UAV assisted dense 5G networks, where reducing delays and traffic offloading are the main objectives. The challenge of delays and transmission delays, particularly in emergencies such as natural disasters, that span vast geographical areas, enabling the use of UAVs as scalable on-demand BSs, the density of which can be adjusted proportionally to the network load. A three-dimensional social tie strength model, which jointly captures social overlap (Jaccard similarity of friend sets), similarity of interests (inverse-popularity-weighted content preferences) and contact quality (Gamma-distributed contact duration probability), governs D2D cluster head (CH) selection and resource block (RB) allocation. A scheme of heuristic resource allocation, HSARA (Heuristic Social-Aware Resource Allocation) is proposed to solve the interference management problems intrinsic in the coexistence of D2D clusters and cellular users sharing a common spectrum; the algorithm proceeds through a deferred-acceptance initialization phase and a swap-matching refinement phase and is proven to converge to a stable bilateral exchange matching in a finite number of steps. Simulations are conducted in MATLAB R2020a for a downlink single-cell UAV-assisted dense network in which users are uniformly distributed within a 500 m radius. The proposed HSARA scheme achieves significant throughput gains and nearly quadruples the performance of the social-aware, social-unaware, and MSARA baseline schemes as network density increases, while a non-trivial optimal UAV altitude of 300 m is identified that balances line-of-sight gain against induced interference.]]></description>
            <category>ARTICLE</category>
        </item>
        <item>
            <title><![CDATA[Multimodal beehive weight forecasting using IoT telemetry and acoustic signal processing]]></title>
            <link>https://reference-global.com/article/10.2478/jee-2026-0044</link>
            <guid isPermaLink="false">/article/10.2478/jee-2026-0044</guid>
            <pubDate>Thu, 27 Aug 2026 00:00:00 GMT</pubDate>
            <description><![CDATA[Accurate forecasting of beehive weight is an important challenge in precision apiculture, as it enables continuous assessment of colony development, nectar-flow dynamics, and potential abnormal events. This paper presents a multimodal Internet of Things (IoT) telemetry framework for short-term beehive weight forecasting using data collected by the Intelligent Hives monitoring system. The proposed framework integrates load-cell measurements, environmental sensing, acoustic telemetry, edge data acquisition, and cloud-based analytics. The dataset contains more than 3.4 million telemetry records acquired from over 250 instrumented beehives operating in real-world apiaries between 2022 and 2025. Several forecasting approaches, including multiple linear regression, gradient boosting, and feedforward neural networks, were evaluated using MSE, RMSE, and R² metrics. The experimental results indicate that historical weight measurements constitute the dominant source of predictive information, while environmental and acoustic telemetry provide complementary contextual information for multimodal hive monitoring. Furthermore, lightweight regression-based models were found to offer a favorable trade-off between forecasting accuracy and computational complexity, supporting their deployment on resource-constrained IoT platforms. The presented results demonstrate the practical applicability of multimodal telemetry, signal processing, and machine learning techniques for intelligent monitoring and predictive analytics in precision agriculture.]]></description>
            <category>ARTICLE</category>
        </item>
        <item>
            <title><![CDATA[Maxwell-Wagner limit of copper contacts on DNTT transistors]]></title>
            <link>https://reference-global.com/article/10.2478/jee-2026-0042</link>
            <guid isPermaLink="false">/article/10.2478/jee-2026-0042</guid>
            <pubDate>Thu, 27 Aug 2026 00:00:00 GMT</pubDate>
            <description><![CDATA[The contact resistance of dinaphtho[2,3-b:2′,3′-f]thieno[3,2-b]thiophene (DNTT) thin-film transistors with copper source/drain electrodes was extracted by the transmission-line method between 200 and 400 K on two gate-dielectric treatments (OTS and HMDS). Resolved interval by interval and at constant gate overdrive, the activation energy of the contact resistance vanishes below 260 K and above 300 K, and all of the activation is concentrated in the crossover between these two plateaus, which yields an injection barrier of 0.63 eV. The contact is therefore limited by a succession of mechanisms rather than by one: tunnelling at low temperature, where the available fields lie one to two orders of magnitude below the Fowler-Nordheim scale so that the barrier remains trapezoidal; thermionic emission through the crossover; and, once injection ceases to limit, a temperature-independent floor of 3-6 kΩ·cm, voltage independent once current crowding is removed, consistent with the Maxwell–Wagner space-charge limit for a DNTT conductivity of about 10−7 S/cm. The conventional plot of activation energy against the square root of field is linear and would alone indicate Schottky barrier lowering, but it fails the thermal-scaling test that barrier lowering equally imposes; current crowding, which accounts for the high-temperature field exponent, cannot generate the crossover. Both surface treatments give the same physics and imply the same conductivity to within a factor of two, indicating a floor set by the dielectric relaxation of the DNTT itself rather than by the metal–organic energetics.]]></description>
            <category>ARTICLE</category>
        </item>
        <item>
            <title><![CDATA[Automatic vineyard health control: A review of existing systems and datasets for disease identification]]></title>
            <link>https://reference-global.com/article/10.2478/jee-2026-0043</link>
            <guid isPermaLink="false">/article/10.2478/jee-2026-0043</guid>
            <pubDate>Thu, 27 Aug 2026 00:00:00 GMT</pubDate>
            <description><![CDATA[In recent years, computer vision and machine learning technologies have enabled new approaches to automatic vineyard health monitoring. This review synthesizes findings from 30 highly relevant studies covering automatic vineyard disease identification systems, examining system architectures ranging from ground-based robots and unmanned aerial vehicles (UAVs) to IoT-integrated platforms, mobile applications, and embedded devices. We analyse the computer vision and machine learning methods employed, including convolutional neural networks for classification, semantic segmentation, object detection, and multimodal sensor fusion techniques. We identify and describe publicly available and custom datasets used for model training and validation, highlighting important differences between controlled laboratory datasets and field-acquired imagery that affect model generalizability. The restricted generalizability is documented by training an EfficientNet-B2 classifier on controlled-background PlantVillage images and testing it on an independent dataset of field-acquired images. Although the model achieved 99.61 ± 0.17% accuracy on the internal test set, its accuracy decreased to 66.80 ± 8.38% and its mean F1-score to 58.33 ± 5.90% on the external field dataset, demonstrating a substantial domain shift between laboratory and vineyard conditions. Our analysis reveals that, while many systems achieve classification accuracy above 90%, significant challenges remain in the early detection of pre-symptomatic infections, robust multi-disease differentiation, and practical deployment across diverse real-world vineyard conditions. We discuss current limitations and outline priority directions for future research, including the need for standardized benchmarks, that require large annotated and balanced datasets, integration of emerging deep learning architectures such as Vision Transformers and foundation models, and development of low-cost, accessible solutions for widespread adoption.]]></description>
            <category>ARTICLE</category>
        </item>
        <item>
            <title><![CDATA[Degradation of Mordant Blue 9 by BDD/TiO2 photoelectrode using 3D printed photoreactor]]></title>
            <link>https://reference-global.com/article/10.2478/jee-2026-0040</link>
            <guid isPermaLink="false">/article/10.2478/jee-2026-0040</guid>
            <pubDate>Thu, 27 Aug 2026 00:00:00 GMT</pubDate>
            <description><![CDATA[This work investigates boron-doped diamond (BDD) / TiO2 heterostructure photoelectrodes for the degradation of the organic dye Mordant Blue 9. The combination of BDD and TiO2 enables enhanced photoelectrocatalytic activity under UV irradiation. The electrodes were characterized using scanning electron microscopy and Raman spectroscopy. Electrochemical behavior was studied using linear sweep voltammetry under UV illumination. The degradation efficiency was evaluated with and without applied bias voltage. Results demonstrate successful degradation of Mordant Blue 9 and a significant improvement in degradation rate when both UV irradiation and external potential are applied. The study confirms the suitability of BDD/TiO2 structures for advanced water treatment applications.]]></description>
            <category>ARTICLE</category>
        </item>
        <item>
            <title><![CDATA[Novel ternary logic gate circuits using CNTFET technology for energy-efficient design]]></title>
            <link>https://reference-global.com/article/10.2478/jee-2026-0036</link>
            <guid isPermaLink="false">/article/10.2478/jee-2026-0036</guid>
            <pubDate>Thu, 27 Aug 2026 00:00:00 GMT</pubDate>
            <description><![CDATA[Ternary logic, which utilizes three distinct logical states instead of the traditional two, is often considered a more effective alternative to binary logic. This increased state capacity can lead to benefits such as reduced chip size, lower power consumption, and faster response time. The special qualities of carbon nanotube field transistors (CNTFETs), namely their ability to alter diameter and achieve the required threshold voltage, are the foundation of STI design. Because of this characteristic, CNTFETs can be used with multiple-Vth design techniques. This paper presents new designs of ternary logic gates using CNTFET technology. The Synopsys HSPICE with 32 nm CNTFET technology is used to model a ternary inverter and universal gates. The proposed standard ternary inverter achieves a 74.45% reduction in power delay product compared to the existing design. Further, the proposed ternary NAND and NOR gates demonstrate a 50.95% and 81.85% reduction in average power consumption, respectively. Finally, the simulation results demonstrate that the proposed gates exhibit a significant improvement in the power-delay product and can be effectively utilized in energy-efficient designs.]]></description>
            <category>ARTICLE</category>
        </item>
        <item>
            <title><![CDATA[An efficient codebook design approach based on salp swarm algorithm for the downlink SCMA system]]></title>
            <link>https://reference-global.com/article/10.2478/jee-2026-0038</link>
            <guid isPermaLink="false">/article/10.2478/jee-2026-0038</guid>
            <pubDate>Thu, 27 Aug 2026 00:00:00 GMT</pubDate>
            <description><![CDATA[The codebooks are the key components of the sparse code multiple access (SCMA) system. In order to distinguish the multiple users from one another, a distinct codebook is assigned to each user. Apart from this, the codebook design has a direct effect on the SCMA performance. In other words, a properly designed codebook set leads to a performance boost in the SCMA system. For this reason, it is crucial to find the most appropriate codebook set to maximize the system performance. With that purpose in mind, a salp swarm algorithm (SSA)-based codebook design scheme has been proposed for the downlink SCMA system in this paper. In the proposed scheme, the codebooks are directly optimized by a recently developed robust algorithm called SSA to acquire the ideal codebook set for the related system. Most of the codebook design procedures available in the literature consider some key performance indicators like minimum product distance (MPD) and minimum Euclidean distance (MED) of the superimposed codewords while designing the codebook set. However, the system performance represented by the average bit error rate (BER) of the multiple users cannot only be determined by the distance properties of the superimposed codewords but also affected by the symbol detector by which the information bits are recovered using the designed codebook set at the receiver side. For this reason, in order to take the whole system into account without concentrating solely on the distance characteristics of the codewords, the average BER value of the SCMA scheme is employed as a fitness function of SSA to be minimized while optimizing the codebook set. In the simulation study, the BER achievement of the proposed SSA-based codebook design strategy is compared to those of varied methods existing in the literature. According to the simulation results, the codebook set produced by the suggested strategy makes the downlink SCMA system to reach better BER levels compared to the other methods considered in this paper.]]></description>
            <category>ARTICLE</category>
        </item>
        <item>
            <title><![CDATA[Uncertainty estimation of receiver antenna gain for UAV-based antenna measurements]]></title>
            <link>https://reference-global.com/article/10.2478/jee-2026-0039</link>
            <guid isPermaLink="false">/article/10.2478/jee-2026-0039</guid>
            <pubDate>Thu, 27 Aug 2026 00:00:00 GMT</pubDate>
            <description><![CDATA[The article deals with the issue of estimating the measurement (receiver) antenna gain uncertainty for Unmanned Aerial Vehicle (UAV)-based radiation pattern measurement of large FM antenna system (AS). The ± 5° tolerance for the axis of UAV rotation (pitch, yaw, roll) is imposed, and the simulation of the measurement device, receiver antenna and UAV is verified by aerial and static semi-anechoic chamber radiation pattern measurements in a horizontal cut at theta = 90°. The simulation model is used to further estimate the gain deviation limited by the imposed rotation criteria, resulting in a measurement antenna gain expanded combined standard uncertainty of 0.44 dB for the given measurement system.]]></description>
            <category>ARTICLE</category>
        </item>
    </channel>
</rss>