A photoplethysmography is a noninvasive heart beat monitoring instrument with optical components. As develop miniaturization and wireless communication, portable health monitoring sensors with reduced size are demanded. Additionally, continuous health monitoring and precaution devices for emergency of vital signals are required for ages. To develop the device under these demands, the photoplethysmography is requested to be minimized in size and to be robust against noise. A major source of noise is body motion; motion artifact reduction is the most significant technique to develop.
In this thesis, a wireless instrument with reduced motion artifact is proposed. The device is in a form of finger band, and measure the bio-signal on the finger base. To reduce the size, components of the device, such as filters and amplifiers, are minimized and optimized. Moreover, this research presents an active noise cancellation method using a 2-D accelerometer to reduce signal distortion by body motion. The accelerometer with photoplethysmography detects the hand acceleration in two easily distorted directions. The adaptive filter is then used to reconstruct the distorted signal which is sensitive to the motion, using the correlation between the accelerometer sensor signals and the distorted photoplethysmograph signal. Normalized LMS (least mean square) algorithm is proposed due to the program size and operation speed for portable device. The filter coefficients are adapted in real time.