<![CDATA[Electrocardiogram (ECG) examination is essential for detecting heart rhythm disorders, yet limited access and high costs often prevent routine medical check-ups for many people. This study addresses these obstacles by designing and developing a portable ECG prototype capable of independent home-based heart monitoring. The system integrates an AD8232 sensor for signal acquisition, an Arduino Uno microcontroller as the main processor, and a simplified Random Forest classification algorithm to distinguish between normal, bradycardia, and tachycardia conditions. Measurement results are saved in CSV format on an SD card, then visualized and analyzed using Jupyter Notebook. The prototype was tested on 100 samples in a static and relaxed state to ensure signal stability. Its heartbeat classification achieved an accuracy of 99.0%, slightly higher than the PTB-XL reference dataset’s 98.0%, and consistent with results reported by recent TinyML- and Random Forest-based ECG studies. Unlike prior IoT-based frameworks, this work combines cost-effective microcontroller hardware with simplified offline on-device classification for practical daily monitoring without continuous cloud access. These findings confirm that the proposed system can produce reliable readings approaching clinical standards while remaining simple, affordable with a component cost under USD 31, and accessible for routine public heart health screening.]]>
Copyrights © 2025
