Publikasi Prosiding

Epilepsy is a neurological disorder that can be identified through changes in the electrical activity of the brain recorded in electroencephalography (EEG) signals. The nonstationary nature of EEG signals requires appropriate feature extraction and classification methods. This study compares the performance of an Artificial Neural Network (ANN) and Naive Bayes in detecting epileptic activity using Discrete Wavelet Transform (DWT) features. EEG data were obtained from four patients in the CHB-MIT Scalp EEG Database, sampled at 256 Hz using 23 channels. The research pipeline comprised preprocessing, 2-s segmentation with 50% overlap, four-level decomposition using the Daubechies-4 wavelet, extraction of 460 features, class balancing, data splitting, standardization, classification, and evaluation. Processing produced 21,594 initial segments, including 335 seizure and 21,259 non-seizure segments. Random undersampling yielded 670 balanced segments, which were divided into 468 training, 100 validation, and 102 testing samples. The ANN achieved an accuracy of 96.078%, sensitivity of 94.545%, specificity of 97.872%, precision of 98.113%, F1-score of 96.296%, and AUC of 0.9957. Naive Bayes achieved an accuracy of 95.098%, sensitivity of 94.545%, specificity of 95.745%, precision of 96.296%, F1-score of 95.413%, and AUC of 0.9756. Naive Bayes required less training time at 0.4826 s, compared with 10.421 s for the ANN. However, the ANN had a shorter testing time and slightly better classification performance. Therefore, the ANN is more suitable when classification performance is the priority, whereas Naive Bayes is preferable when training efficiency is more important. These findings remain limited to segment-level evaluation and do not demonstrate cross-patient generalization.

Single-phase induction motors serve as prime movers in various industrial applications, necessitating condition monitoring systems capable of early fault detection. This research aims to develop an ESP32-based data processing and decision-making system for electric motor condition monitoring on the MOTGUARD platform. The system utilizes DS18B20 temperature, ACS712 current, and ADXL345 vibration sensors to acquire real-time motor condition data. Sensor data is processed using the Exponential Moving Average (EMA) method to reduce noise, the Root Mean Square (RMS) method to obtain effective AC current and vibration values, and an adaptive baseline method to establish the normal vibration state for each motor. Decision-making employs OR logic based on threshold values of 80°C for temperature, 4 A for current, and 2 m/s² for vibration. Testing results demonstrate that all sensors accurately measure temperature, current, and vibration parameters as designed. The system successfully identifies normal and alarm conditions in real-time and displays motor status information on the monitoring interface. The findings indicate that MOTGUARD enhances the reliability of single-phase induction motor condition monitoring and provides early warnings regarding potential faults.

Fenomena komodifikasi simbol agama di era digital telah menggeser fungsi bahasa Arab dari ranah sakral menjadi instrumen pasar yang bernilai ekonomi. Penelitian ini bertujuan untuk menganalisis secara mendalam pergeseran semantik dan pragmatik frasa bahasa Arab yang diadopsi oleh media gaya hidup kontemporer di Indonesia. Dengan menggunakan metode kualitatif deskriptif melalui pendekatan analisis wacana kritis (CDA) dan sosiolinguistik, data berupa frasa populer seperti hijrah, syar'i, dan muhasabah dikumpulkan dari berbagai platform media sosial dan majalah digital sepanjang tahun 2025 hingga 2026. Hasil penelitian menunjukkan adanya gejala desakralisasi dan rekontekstualisasi bahasa Arab. Makna spiritual yang melekat pada kosakata tersebut mengalami reduksi demi kepentingan komersial dan pembentukan citra identitas kelas sosial Muslim urban. Bahasa Arab tidak lagi sekadar berfungsi sebagai alat komunikasi teologis, melainkan telah bertransformasi menjadi bentuk symbolic branding dalam industri halal modern. Implikasi akademis penelitian ini sangat signifikan bagi peninjauan kurikulum Program Studi Pendidikan Bahasa Arab (PBA), khususnya dalam memperkaya materi kuliah sosiolinguistik Arab yang kontekstual, sehingga mahasiswa memiliki daya kritis yang kuat terhadap perkembangan bahasa di ruang publik digital.

The utilization of solar photovoltaic (PV) systems in public facilities, such as gazebos, has continued to increase to support sustainable energy provision. However, most existing systems focus either on energy monitoring or access control independently, limiting their capability to provide integrated facility management. This study proposes and implements an Internet of Things (IoT)-based Smart Access Gazebo Energy Surveillance system using an ESP32 microcontroller as the central integrated controller. The proposed system incorporates a PZEM-017 sensor for DC electrical parameter monitoring, a DHT22 sensor for temperature and humidity measurement, an LM393-based LDR module for ambient light detection, and Blynk and a 16 × 2 LCD for both remote and local monitoring. System performance was evaluated through sensor accuracy validation against reference instruments and simultaneous operational stability testing. Experimental results demonstrate that the PZEM-017 sensor achieved mean measurement errors of 0.94% for voltage, 3.90% for current, and 4.86% for power, all of which are within the manufacturer's specified tolerance limits. The DHT22 sensor yielded root mean square errors (RMSEs) of 0.18 °C for temperature and 0.45% RH for relative humidity, while the LM393-based LDR module successfully distinguished bright and dark conditions with 100% classification accuracy. Furthermore, the ESP32 reliably performed concurrent sensor data acquisition, LCD display updates, cloud communication, and automated control without operational interruption throughout the testing period. The findings demonstrate that integrating energy monitoring, environmental monitoring, and access control into a single low-cost ESP32-based platform significantly improves the operational efficiency, security, and reliability of PV-powered gazebo systems. The proposed architecture offers a practical and scalable solution for smart renewable-energy-powered public facilities.

Conventional volumetric hot-mix asphalt design is widely used to determine optimum asphalt content, yet it does not always guarantee a balanced resistance to rutting, fatigue cracking, thermal cracking, and moisture-related distress. This systematic literature review examines the integration of Balanced Mix Design (BMD) with dynamic load modeling for hot-mix asphalt mixtures. The revised manuscript expands the citation corpus with 56 journal articles published in 2022-2026, grouped into four clusters: BMD and performance validation, dynamic modulus and viscoelasticity, recycled/modified materials, and machine learning or optimization. The synthesis indicates that BMD is most useful when it bridges volumetric design, laboratory performance testing, and mechanistic modeling of pavement response under repeated traffic loading. The review proposes an integrated BMD-dynamic loading framework for future asphalt mixture specifications in Indonesia.

The development of simple automation technology has encouraged the implementation of more efficient and easily controllable lighting systems. One of its applications is smart lighting, which is capable of automatically turning lights on and off based on environmental light intensity conditions. This project aims to design and construct a smart lighting prototype for a miniature house using an Arduino Uno, an LDR sensor, a 5V relay, a 12V adapter, and parallel-connected LED lights. The system works by utilizing the LDR sensor as a light detector. Data from the sensor is read by the Arduino Uno through analog pin A0. When the environmental condition is dark, the Arduino sends a signal to the relay through digital pin D8, activating the relay and turning on the LED lights. Conversely, when the condition is bright, the relay is deactivated and the LED lights turn off. Digital pin D2 is used as an optional connection from the digital output of the LDR sensor. The project methodology includes design planning, wiring diagram creation, preparation of tools and materials, miniature house construction, assembly of the control and electrical systems, and functional testing. The test results show that the system operates as designed, with the lights turning on in dark conditions and turning off in bright conditions. Thus, this prototype can serve as an effective learning medium for understanding the working principles of light sensors, microcontrollers, relays, and automatic lighting systems.