Gait patterns classification using spectral features

Accelerometry has been shown to be a good tool for ambulatory activity monitoring. This paper describes the use of spectral features for classification of gait activities based on accelerometric data. The classification is performed by a Gaussian mixture model (GMM) based statistical classifier at the back end. Fifty subjects participated in the experiment and an overall classification accuracy of 86% was achieved using the proposed 25 dimensional features for five different human gait patterns including walking on level surfaces, walking up and down stairs and walking up and down ramps.