This study evaluates the respiratory rate from the signal
of a single-axis accelerometer fixed at the top of the abdomen and for
three types of breathing: slow, normal and fast. Different filters are
applied and compared to determine the most accurate method to
extract this information. The results showed that the use of a third-
order low pass Butterworth filter following an initial estimate of the
respiratory rate was the most accurate method. This study
demonstrates the potential of the accelerometer as a low-cost, non-
intrusive method of screening for sleep disorders and patients follow-
up