本研究的目的為降低應用於穿戴式裝置的心率偵測系統因動作雜訊引起的偵測誤差；此系統的心率訊號是使用光體積描記術(Photoplethysmography, PPG)從人體感測而得到的。本研究所設計的PPG感測系統包含PWM調變器、紅外線LED、光敏電阻 (LDR) 、 PWM解調器、高通濾波器、低通濾波器、和增益放大電路。此外，本研究利用三軸加速度感測器來取得PPG感測器位置的人體動作訊號，其輸出的訊號可能與心率訊號裡的動作雜訊有相當大的關聯性。加速度感測器之輸出訊號與PPG訊號被取樣並轉換為數位訊號後輸入至個人電腦內，使用MATLAB程式進行心率的偵測以及降低動作雜訊對心率偵測的干擾。主動噪音消除(ANC)演算法被用來還原因受到動作雜訊影響而造成失真的心率訊號，其中加速度感測器之輸出訊號是ANC的參考訊號。本研究之ANC演算法的性能經過實驗驗證，比較的標準是市售的一款心率帶。本研究設計的系統在有小動作干擾時的心率誤差為3.52%，在有大動作干擾時為8.81%，而在1 Hz的動作干擾下為4.32%。本研究證明了主動噪音消除法適合用來降低PPG心率訊號裡的動作雜訊，即使動作雜訊的頻率接近心跳的頻率亦然。;This research was aimed at reducing the error caused by motion artifacts in a heart rate detection system to be applied in wearable devices. The heart rate signal was obtained by using the photoplethysmography (PPG) sensor. The PPG sensor we designed consists of a PWM modulator, an infrared LED, a light dependent resistor (LDR), a demodulator, a high-pass filter, a low-pass filter, and an amplification circuit. Furthermore, a 3-axis accelerometer sensor was used to sense the body motion at the PPG sensor site. The output of the accelerometer sensor may have a considerable correlation with the motion artifact in the heart rate signal. Both the PPG heart rate signal and the accelerometer output signal were sampled and digitized through a data acquisition system of a personal computer. The reduction of motion artifact and the heart rate detection were conducted by the MATLAB program in the computer. Taking the accelerometer output signal as the reference signal, an active noise cancellation (ANC) algorithm recovered the corrupted heart rate signal from motion artifact. The performance of the active noise cancellation is evaluated using a commercial heart-rate belt as the golden standard. The heart rate detection error of our system is 3.52% with a small motion, 8.81% with a big motion, and 4.32% with a 1-Hz motion. The result of our experiments proves that the active noise cancellation method is suitable for removing motion artifact from heart rate signal even if there is overlapping between the spectra of the motion artifact and the heart rate signal.
