SLEEP APNEA DETECTION DEVICE USING ECG SIGNALS

• Main Author: Devindra Adara, Fitya
• Format: Final Project
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/84870
• Institution: Institut Teknologi Bandung

Sleep apnea is a sleep disorder that causes individuals to stop breathing for 10 seconds or more while sleeping. If left untreated, sleep apnea can increase the risk of cardiovascular diseases and metabolic disorders. Sleep apnea can be diagnosed using a device called Polysomnography (PSG). However, this device uses numerous sensors, making it uncomfortable for patients. Additionally, it can only be used in hospitals and is quite expensive. Therefore, in this final project, a simple device to detect sleep apnea using only ECG signals has been designed. This device is made to be wearable, making it easier and more practical to use compared to PSG. It uses an AD8232 1-lead ECG sensor and an ESP32 microcontroller. The device can classify conditions of apnea and normal breathing using a deep learning algorithm. The classification results are displayed on a website, showing the time of apnea occurrences and a classification graph in one recording. Among several deep learning algorithms tested, the Apneanet algorithm showed the best accuracy and was therefore chosen to be implemented in the device. Apnea classification is conducted every minute, and the results can be viewed on the website. The accuracy of the Apneanet model on the dataset is 98.77%, meaning this model can effectively distinguish between apnea and normal conditions on the test data derived from the dataset. The model was also tested on real data collected directly from the device, but the data lacked labels. Therefore, the Apneanet model was compared with other existing models, showing a classification result similarity of 84.37%.