The increasing demand for high-performance wireless communication systems has driven the development of compact antennas that can support wideband operation, particularly in the 5 GHz frequency range. Microstrip patch antennas (MPAs) remain attractive for such applications due to their low profile and ease of integration, although achieving a wide impedance bandwidth remains a key challenge. This paper presents a low-profile MPA with a wideband response designed for the 5 GHz band of wireless communication systems. The proposed MPA is implemented on a single-layer FR-4 substrate with a thickness of 1.6 mm and is excited using a coaxial probe feed. To achieve a wide impedance bandwidth, a U-shaped patch is utilized as the primary radiator, complemented by a slot, a parasitic element, and a rectangular ring surrounding the main resonator, aiming to generate multiple resonant modes and improve impedance matching. The influence of each modification on antenna performance is systematically analyzed. To validate its performance, the proposed MPA is fabricated and measured, with the obtained results compared to the simulated ones. Simulation and measurement results show good agreement, confirming that the proposed MPA achieves a wideband response with a fractional bandwidth of 18.8% and a peak realized gain of 4.85 dBi. These results highlight the effectiveness of the proposed design and demonstrate its potential for compact and wideband wireless communication devices operating around 5 GHz frequency range.

Skin diseases are among the most common health issues in domestic cats. However, access to veterinarians is often limited, especially in low-resource settings. Automated image-based detection offers a fast and affordable alternative for early intervention. This paper presents a lightweight approach for diagnosing feline skin diseases using EfficientNetV2 optimized for low-resource devices. A balanced custom dataset consisting of 720 images across nine classes, namely Healthy, Mild/Severe Ringworm, Mild/Severe Acne, Mild/Severe Flea, and Mild/Severe Scabies, was compiled from Kaggle, Roboflow, and Google Images, ensuring ethical use of publicly available data. The images were augmented through rotations (0°, 90°, 180°, 270°) and horizontal flips, resulting in 5,760 images, to enhance model generalization. Five CNN architectures were benchmarked: DenseNet121, MobileNetV2, MobileNetV3, EfficientNetB0, and EfficientNetV2B0. Training was conducted with grid searches over batch sizes {64, 32, 16, 8} and learning rates {1e-3, 5e-4, 2e-4, 1e-4, 5e-5} for up to 300 epochs, and with the Adam optimizer and Reduce-LR-on-Plateau (decay factor 0.5). Early stopping (patience = 10) was used to mitigate overfitting. The best model was selected based on highest validation accuracy. The experiments were conducted on an Intel Xeon 6 CPU (2.2 GHz, 2 vCPUs) in Google Colab without GPU to simulate low-resource deployment. EfficientNetV2B0 achieved the best performance with 99.62% validation accuracy and 99.79% test accuracy, with an average inference latency of 78 ms/frame. Compared to previous studies focusing on heavyweight models or conventional ML using handcrafted features, this work highlights the feasibility of deploying an accurate real-time diagnostic pipeline on edge devices.

Accurate classification of chest X-ray (CXR) images is vital for early detection of thoracic diseases such as COVID-19, Tuberculosis, and Pneumonia, particularly in regions with limited radiological expertise. While deep learning has shown promise in CXR interpretation, many existing models rely solely on internal datasets, risking overfitting and poor generalizability. Furthermore, inadequate tuning of network architectures may limit robustness across varied imaging conditions. This study presents an externally validated deep learning framework based on Convolutional Neural Networks (CNNs) for multi-disease CXR classification. This study compared a baseline CNN with two convolutional layers against a tuned architecture with three layers across multiple image resolutions (64×64, 112×112, 224×224). The proposed model employs transfer learning with a pre-trained CNN, fine-tuned for four-class classification using a softmax output layer. Training was performed with the Adam optimizer (learning rate: 0.0001, batch size: 32) and categorical cross-entropy loss, for up to 50 epochs with early stopping. Internal validation showed the tuned model outperformed the baseline, achieving 0.97 accuracy and an F1-score of 0.89. External validation confirmed superior generalizability, with the tuned model attaining an F1-score of 0.83 and an AUC of 0.97 at 112×112 resolution, compared to the baseline’s F1-score of 0.79 and AUC of 0.94. These results highlight the potential of optimized CNN architectures as reliable, scalable tools for radiological decision support in resource-limited healthcare systems. Future work will incorporate explainable AI methods and real-world clinical validation to ensure safe, interpretable deployment.

This research aims to analyze the comparison of power output between Polycrystalline solar panels with and without the use of mirror reflectors. The growing demand for sustainable energy sources necessitates ongoing efforts to enhance the efficiency of photovoltaic (PV) technology. In this study, a comparative experimental design was implemented over a two-day period, utilizing two identical solar panel configurations: a Reflector Equipped-Solar Panel (RESP) and a Solar Panel Without Reflector (WRSP). The electrical parameters of both configurations, including voltage, current, and power output, along with temperature, were continuously monitored and analyzed. The results indicate that the RESP consistently outperformed the WRSP. On average, the RESP produced 28–35% higher electrical current, which directly translated into a 25–32% increase in power output compared to the WRSP. The peak power of the RESP reached 21.3 W, whereas the WRSP peaked at 16.2 W. Although the RESP operated at a higher temperature (approximately 5–7 °C greater than WRSP), the power gain from increased irradiance effectively outweighed the thermal losses. These findings provide strong empirical evidence that reflective augmentation is a viable and cost-effective method for enhancing the performance of standard solar panels, offering valuable insights for maximizing renewable energy harvesting.

The miniaturization of radio frequency (RF) and microwave filters is crucial in modern communication systems, particularly in space-constrained environments such as portable, wearable, or integrated multi-band platforms. In this study, a compact single band-pass filter (BPF) is proposed using an L-shaped resonator loaded with a via-through hole. The incorporation of the via allows the creation of a quarter-wavelength resonator, which significantly reduces the physical size to 44.6% compared to traditional half-wavelength resonators commonly used in band-pass filters. Additionally, the via contributes to an improved impedance bandwidth. Two L-shaped resonators are arranged in a back-to-back configuration to realize a second-order filter response. The structure is excited through a stub-loaded transmission line that couples the signal from port one to port two. To validate the proposed design, the filter was fabricated and measured. Experimental results demonstrate a -3 dB fractional bandwidth (FBW) exceeding 25% at the centre frequency of 3.2 GHz, with a measured insertion loss of -1.662 dB across the passband, indicating good performance for compact filter applications.

The advancement of Internet of Things (IoT) and fog computing technologies has created significant opportunities for more efficient, faster, and proximity-based data management. However, IoT-Fog systems face considerable challenges related to device heterogeneity, traffic dynamics, and the complexity of network topologies that continuously change. This study conducts a Systematic Literature Review (SLR) of various research works covering dynamic scheduling, routing, context-aware data flow, offloading, and IoT-Fog systems without adaptive mechanisms. The findings indicate that most existing approaches still rely on relatively static topology assumptions, rendering them insufficiently adaptive to real-time changes in network conditions. One area identified as a research gap is dynamic frequency control, an adaptive mechanism capable of dynamically adjusting data transmission intensity based on network conditions. The main conclusion of this study emphasizes the necessity for developing adaptive systems that are topology-agnostic and supported by dynamic frequency control to maintain optimal performance even under significant topology changes. Such systems are anticipated to become a crucial foundation for future IoT-Fog applications, including smart cities, Industry 4.0, and intelligent healthcare services, which demand high reliability and low latency.

The deployment of 5G networks in provincial capitals such as Semarang presents challenges related to high user demand, coverage requirements, and investment costs. This research aims to evaluate the network coverage and techno-economic feasibility of 5G implementation in Semarang. Using an urban macro propagation model, two scenarios were analysed: Uplink Non-Line-of-Sight (UL-NLOS) and Downlink Non-Line-of-Sight (DL-NLOS). The UL-NLOS scenario requires 11 sites, while the DL-NLOS scenario requires 6 sites to achieve full coverage of the city. The average Synchronization Signal–Reference Signal Received Power (SS-RSRP) is −125.74 dBm, indicating sufficient signal strength. A techno-economic analysis reveals that the UL-NLOS scheme yields an NPV of IDR 292.566.473.678 with an IRR of 110.25%, while the DL-NLOS scheme yields an NPV of IDR 300.000.000.000 with an IRR of 115.38%. These results confirm that both scenarios are technically feasible and economically viable. The findings suggest that 5G deployment in Semarang can yield profitable returns, providing valuable insights for mobile operators in planning future investments.

This study aims to mitigate the impact of electromagnetic interference on the performance of electronic systems in Unmanned Aerial Vehicles (UAVs) by employing various shielding materials. The materials tested include carbon fiber, E-glass, E-glass with an aluminium foil, and E-glass with a copper foil. A Vector Network Analyzer (VNA) and scattering parameter (S-Parameter) analysis, including reflection, absorption, and multiple reflection, were used to evaluate the shielding effectiveness of these materials within the frequency range of 4 - 5 GHz. The results showed that E-glass coated with copper had good overall shielding performance for SER, SEA, and SET values. This material reached a SET value of 96 dB at a frequency of 4.6 GHz, followed by E-glass coated with aluminium. In addition, adding carbon layers increased the shielding effectiveness, while E-glass without coating had the lowest shielding performance compared to the other materials. These findings indicate that E-glass coated with metal provides superior shielding effectiveness compared to carbon fiber, even when used in greater thickness.

Teknologi terus berkembang dalam menciptakan peralatan listrik yang menunjang kebutuhan manusia seperti fotovoltaik (PV) dan motor listrik DC namun peralatan tersebut membutuhkan pengaturan tegangan yang sesuai agar kualitas pada alat tersebut tetap optimal dan mencegah terjadinya kerusakan.Buck converter merupakan salah satu komponen penting dalam mengatur tegangan.Buck converter memiliki tantangan berupa tegangan yang berubah sehingga kontrol PI merupakan salah satu solusi dalam mencegah terjadinya perubahan tegangan.Parameter pada kontrol PI dioptimalkan menggunakan algoritme PSO yang terkenal dalam memperoleh solusi optimal dengan kecepatan terbilang singkat pada optimasi parameter. Simulasi menggunakan perangkat lunak MATLAB/Simulink untuk mengevaluasi kinerja sistem sebelum dan setelah optimasi. Parameter kinerja diukur dalam membandingkan nilai overshoot dan settling time pada tuning Ziegler Nichols dengan kontrol PI-PSO. Hasil simulasi menunjukkan bahwa penggunaan PSO berhasil meningkatkan kinerja kontrol PI dengan mempercepat settling time dari 6,3 ms menjadi 4,9 ms,. Penelitian ini menunjukkan bahwa kontrol PI yang dioptimalkan dengan PSO memberikan performa kecepatan respon yang lebih unggul daripada tuning PI Ziegler Nichols serta memberikan performa yang stabil dan andal saat mengalami gangguan pada sistem.