Design of Wearable Pulse Oximeter Based on Android with Bluetooth Low Energy Communication Using Artificial Neural Network

A wearable pulse oximeter was created that can monitor a person's oxygen saturation and heart rate at any time. By designing the smallest possible size device, this device can be used even during activities. Users can easily monitor oxygen saturation (SpO2) and heart rate (BPM) on the Android application. For data transmission, the device is supported by Bluetooth Low Energy (BLE) so that the battery used can last longer. With the MAX30100 sensor as a pulse oximeter sensor and nRF52832 as a microprocessor as well as a communication module, the device has been successfully made with dimensions of 5.3 cm x 4 cm x 3 cm that can be used in the user's hand. All SpO2 and BPM detection data were successfully sent to Android via Bluetooth Low Energy (BLE). The device succeeded in detecting the index finger with an average accuracy of Sp02 and BPM of 99.59% and 97.14%, and succeeded in detecting under normal conditions, after waking up and exercising with the largest average Sp02 accuracy of 99.72 % and BPM 98.34 %. Analysis using ANN on Jupyter Notebook 6.0.3 succeeded in reducing detections spikes with an average accuracy of 86.45%. And the device can work at low power (low energy) ie 78.53 mW -107.95 mW with a battery life of 8 hours 33 minutes.