Abstract
Pulse rates obtained from photoplethysmography (PPG) sensors are important for monitoring cardiovascular condition, especially during exercise, because it is a simple, low-cost, and easy to use. However, it is difficult to precisely count pulse rates during exercise because PPG is sensitive to body movement. When a motion artifact occurs, changes in vascular volume cause noise. Differences in cutaneous blood supply to different anatomic regions are apparent from the literature. Here, we consider motion artifacts as related to anatomical measurement sites and light sources. In order to eliminate the motion artifacts, we compared the signal obtained from green PPG to that from infrared PPG with related to different anatomical sites during body movement. PPG were collected simultaneously from three anatomical sites: upper arm, wrist and finger. Eleven young healthy subjects without any indication of peripheral arterial disease were recruited to participate in the experiment. To produce artifacts, the subjects were asked to walk at a pace of 4 km/hr on the treadmill. PPG magnitude and intervals were calculated by peak detection using predetermined thresholds, and the error rate between the PPG pulse rate and the heart rate was then compared. As the results, the magnitude and variability of the green PPG have different nature to those of the infrared PPG. In a resting state, the maximum magnitude of green PPG was located at upper arm, but that of infrared was located at wrist. During walking, all site of infrared PPG were more sensitive to motion artifact than those of green PPG. These results suggested that green PPG reflected different vascular distribution to infrared PPG. Additionally, the green PPG from the upper arm showed less error rate than did the peripheral one. Therefore, the green PPG located at the upper arm may be a suitable for pulse rate monitoring during walking.
