| Summary: | 碩士 === 國立臺灣科技大學 === 電子工程系 === 102 === Exercise is important for our health, however inappropriate exercise would harm or affect nothing to our body. Therefore, a wearable exercise intensity monitoring device is needed to assist user for managing their exercise intensity. One of many parameter index to indicate the exercise intensity is heart rate (HR).
In this thesis, we proposed a high accuracy HR detection method and implemented it on a wearable and low power device for exercise intensity monitoring. The accuracy of HR detection method has been verified into two levels, peak detection and HR detection. Peak detection is used for examining the accuracy of proposed algorithm compared with our implemented Pan Tompkins algorithm. We use MIT BIH ST-Change Database and MATLAB to verify the accuracy of the algorithm. The accuracy of peak detection is 99.2%. For HR detection, we verified the accuracy into three conditions i.e. initial, resting, and dynamic condition. In initial condition, the results show that the algorithm can detect the HR accurately in normal ECG, inverted ECG, and flat waveform. For resting condition, we use a commercial ECG simulator as signal input and obtain 100% for the accuracy of HR detection. For dynamic condition, we use treadmill test with ten subjects (8 male, 2 female) that has been asked to walk with six different speeds ranging from 1.8 km/h to 6.3 km/h and to run with speed 7.2 km/h. The accuracy of HR detection is 99.7%.
|
|---|
