Date Approved

4-2012

Graduate Degree Type

Thesis

Degree Name

Engineering (M.S.E.)

Degree Program

School of Engineering

Abstract

Motion artifact strongly corrupts heart rate measurements in current pulse oximetry systems. In many, almost any motion will greatly diminish the system’s ability to extract a reliable heart rate. The artifact is most likely present due to normally non-pulsatile components of the body, such as venous blood and tissue fluid, which become pulsatile during motion. This paper presents a motion artifact reduction method using an accelerometer that attempts to recover a usable heart rate sensor signal that has been corrupted by motion. The method was developed for a wrist pulse oximeter sensor and was adapted for a ring sensor, both of which were very susceptible to arm motion. An accelerometer was paired with the pulse oximeter to detect the motion. This motion signal was then used to recover the corrupted heart rate signal. The correlation between the acceleration and the heart rate signals was analyzed and two adaptive filter models were created to relate the corrupted signal to the acceleration. These filters were partially successful in removing the motion artifact. The results show that the wrist sensor was much more susceptible to motion in any direction, while the ring sensor was mainly susceptible to motion in the same direction as the digital artery.

Download

Share

COinS