Stroke is a leading cause of death and disability in Indonesia, while early detection is still limited by invasive, expensive, and impractical methods. This study developed a non-invasive stroke detection system based on the MAX30105 sensor, utilizing photoplethysmography (PPG) to measure blood pressure, blood sugar, and cholesterol levels. Data were calibrated using a multi-layer perceptron (MLP) and classified using logistic regression into “Yes/No” stroke risk. The ESP32-based system is integrated with IoT with a real-time display on the LCD and an Android application. Measurement of blood sugar levels using the MAX30105 sensor yielded an accuracy level of 86.79%, while cholesterol measurements achieved 95.07%, systolic blood pressure reached 92.75%, and diastolic blood pressure achieved 97.24%. Additionally, the precision level of the device is indicated by a coefficient of variation value below 2% for all measurement parameters, demonstrating stable and consistent results. The results of the stroke risk classification test obtained an accuracy of 85.71%. The system demonstrated good, consistent performance and has the potential to be a practical solution for non-invasive health monitoring and early stroke detection.
The application of Internet of Things (IoT) technology in modern agriculture has become increasingly important for improving efficiency, productivity, and environmental monitoring. This study presents the design and implementation of an ESP32-based IoT monitoring system for a rice seedling greenhouse. The proposed system was developed to monitor multiple environmental parameters simultaneously, including temperature, soil moisture, pH, and Total Dissolved Solids (TDS) of the nutrient solution. The research employed a Research and Development (R&D) approach involving system requirement analysis, hardware and software design, implementation, IoT integration, and performance evaluation. The hardware architecture consists of an ESP32 microcontroller, DS18B20 temperature sensor, soil moisture sensor, pH sensor, TDS sensor, and an I2C LCD for local data visualization. Sensor data are transmitted via WiFi using the HTTP protocol and displayed in real time through a web-based monitoring platform. Experimental results demonstrated that the developed system successfully acquired, processed, and transmitted environmental data continuously to the monitoring server. Analysis of the collected data showed soil moisture values ranging from 47.8% to 59.1%, temperature values between 28.13°C and 31.69°C, TDS values from 118.22 ppm to 138.02 ppm, and pH values between 3.27 and 13.78. The results also revealed an inverse relationship between temperature and soil moisture, indicating the influence of environmental temperature on water evaporation within the growing medium. Furthermore, the monitoring platform enabled real-time remote supervision and historical data storage, supporting data-driven decision-making for greenhouse management. Overall, the proposed system demonstrates the effectiveness of integrating electronic sensing devices and IoT technology to support smart agriculture applications in rice seedling cultivation.
Abstract − Conventional Maximum Power Point Tracking (MPPT) methods in photovoltaic (PV) systems frequently suffer from significant efficiency degradation when subjected to dynamic weather and Partial Shading Conditions (PSC). To address this issue, this study proposes a novel hybrid control algorithm integrating Bayesian Optimization (BO) and Deep Reinforcement Learning (DRL). The primary contribution of this research is the development of an adaptive MPPT system architecture that leverages the global exploration capabilities of BO alongside the high-precision local tuning of DRL to maximize solar energy extraction. The methodology evaluates the proposed BO-DRL agent through an ablation study within a Python simulation environment across four distinctive environmental profiles: uniform irradiance, light partial shading, heavy partial shading, and extreme dynamic conditions. In this framework, the BO component executes a probabilistic global search via Gaussian Processes to prevent the system from getting trapped in local maxima, while the DRL agent performs continuous duty cycle adjustments to minimize steady-state oscillations. Simulation results demonstrate that the hybrid approach significantly outperforms the conventional Perturb and Observe (P&O) method. Under heavy partial shading, the hybrid algorithm achieves a tracking efficiency of 96.08%, whereas the P&O method drops to 62.24% due to local peak entrapment. Under extreme dynamic scenarios, the hybrid efficiency remains robust at 93.22%, while the P&O performance drastically degrades to 39.07%. Furthermore, the ablation validation proves that standalone DRL agents fail to initialize optimally without the global search assistance from the BO unit. In conclusion, the synergistic integration of BO-DRL yields a highly robust, efficient, and adaptive MPPT control solution capable of optimizing PV energy harvesting in highly volatile environments.
This study investigates the use of Singular Value Decomposition (SVD) as an image compression technique to improve the efficiency of deep learning models for face mask detection. The proposed approach applies SVD-based compression with different values of k (k = 10, 30, 50) prior to training a MobileNetV2 model. Experimental results show that SVD-based compression can significantly reduce computational cost while maintaining high classification performance. The model trained with k = 50 achieves the highest accuracy of 99.43%, slightly outperforming the model trained on the original dataset. In addition, compressed datasets require less training time, with the fastest configuration (k = 30) achieving a substantial reduction in training duration. The results also indicate that moderate compression levels provide an optimal balance between efficiency and accuracy, while excessive compression leads to performance degradation due to loss of important image features. Furthermore, the training process demonstrates faster convergence for compressed datasets, indicating improved learning efficiency. Overall, this study confirms that SVD-based image compression is an effective preprocessing technique for improving deep learning efficiency without significantly compromising classification accuracy.
The growing demand for renewable energy drives the need for efficient and reliable power conversion systems. This paper presents the simulation and modeling of a CUK H-Bridge Inverter with current control using a Proportional Integral (PI) algorithm implemented in Power Simulator (PSIM) software. The proposed topology combines an H-Bridge inverter with a CUK AC–AC converter, enabling bidirectional power flow, buck-boost voltage operation, and low Total Harmonic Distortion (THD) in the output current. The current control loop uses a PI controller tuned using the empirical Ziegler–Nichols (Z-N) method, followed by fine-tuning, resulting in optimal parameters of Kp = 0.6, Ti = 0.0167, and Ki = 35. Simulation results show that the actual output current (Iact) accurately follows the reference current (Iref) with minimal steady-state error. The topology operates in buck mode (Vo < Vin), steady-state mode (Vo ≈ Vin), and boost mode (Vo > Vin) depending on the given reference current value. A maximum output power of 2.25 kW is achieved at a reference current of 15 A with a maximum output voltage of 150 V. The measured output current THD is 3.6%, meeting the IEEE 519 standard. These results confirm the effectiveness of the PI-controlled CUK H-Bridge Inverter topology for renewable energy applications.
