Classification of walking pattern using acceleration waveform in elderly people

We attempted to classify walking on level ground from walking on a stairway using a waist acceleration signal. A tri-axial accelerometer was fixed to the subject's waist and the three acceleration signals were recorded by a portable data logger at a sampling rate of 256 Hz. Eleven healthy, elderly subjects were asked to walk through a corridor and up and down a stairway as a single sequence, without any instruction. The data were analyzed using a discrete wavelet transform. Walking patterns were classified using two parameters; one was the ratio between the power of wavelet coefficients which were corresponded to locomotion and total power in the anteroposterior direction (RPA). The other was the ratio between root mean square of wavelet coefficients at the anteroposterior direction and that at the vertical direction (RAV). Walking up stairs could be distinguished by the smallest value in RPA from other walking patterns. Walking down stairs could be discriminated from level walking using RAV. It was possible to classify the walking pattern using acceleration signals in elderly people.