HARDWARE AND PROCESSING ALGORITHM DESIGN FOR PHOTOPLETHYSMOGRAPHY-BASED WEARABLE HEART RATE MONITORING DEVICE

• Main Author: Gede Indrayana Yogaputra, I
• Format: Final Project
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/66578
• Institution: Institut Teknologi Bandung

Cardiovascular disease is one of the main causes of death not only in Indonesia, but also worldwide. Regular monitoring of the heart rhythm and its abnormalities is necessary in indicating the possibility of early cardiovascular disease. Understanding the limited public access to conventional heart rate monitoring based on ECG, we propose a wearable, photoplethysmography-based heart rate monitoring device which aims to provide a monitoring system that is accessible, easy to use in everyday life, and could be used without the help of medical personnel. Photoplethysmography is an optical method for obtaining heart rate signals using LED and photodetector; the difference in the intensity of the LED light reflected to the photodetector carries information about the volume of blood flowing under the skin. The proposed device works with Android-based computing concepts, where wearables only work to transmit data wirelessly to Android, while Android devices will perform heart rate estimation. The key feature, as well as novelty point provided by the design is the ability to detect tachycardia (resting heart rate above normal threshold) and tachycardia (resting heart rate above normal threshold) through calculation of average heart rate and RNN-LSTM based activity classification. The design process includes the design of wearable device and Android applications which serves as user interface. These two main sections are designed through a top-down approach and are divided into four subsystems, including the Data Acquisition subsystem, the User I/O & Control subsystem, the Data Processing subsystem, and the Graphic User Interface subsystem. This book discusses in more detail about the selection of hardware components and the selection of heart rate algorithm, which are the responsibility of the author. The hardware component consist of a PPG Gravity sensor, an MPU6050 accelerometer sensor as a motion reference, a power path circuit, an ESP32 controller, two LEDs, an MSS22D18 switch, and a TP4056 charging module which is realized into three layers of PCB measuring 44 ? 38 mm. To estimate heart rate, the Galli algorithm based on Subspace Decomposition and Kalman Filtering is used; this algorithm is proven to have the highest accuracy among the six algorithms tested by the team. The device managed to achieve five of the six specifications targeted: covering a heart rate range of 30 – 250 bpm, AAE (Average Absolute Error) accuracy < 6 bpm iv (device reached an accuracy rate of 3,646 bpm) with a resolution of 1 bpm, sampling rate > 50 Hz (the device reaches 59 Hz), the estimated output rate is 0.5 – 1.5 Hz (reaches 0.988 Hz), and the maximum distance between the wearable and Android device is > 7 meters (reaches 40 meters). The product reaches length, width, and mass (< 50? 40 mm, < 200 gram) according to specifications, but is 3 mm higher from the maximum specification of 20 mm; this can be improved in the future by reducing the use of breakout modules in the design. The device successfully solves the main problem, by providing an alternative to heart rate monitoring that is easy to use due to its shape and size, is accessible to the public, can be used without medical assistance, and is proven to be accurate in various activities. Regarding development, product accuracy can be improved in the future through the use of sensors with higher signal-to-noise ratio and consistent signal quality in moving conditions, while battery life can be increased by selecting controllers that require less current than the ESP32.