Gait phase detection and discrimination between walking–jogging activities using hidden Markov models applied to foot motion data from a gyroscope

In this paper we present a classifier based on a hidden Markov model (HMM) that was applied to a gait treadmill dataset for gait phase detection and walking/jogging discrimination. The gait events foot strike, foot flat, heel off, toe off were detected using a uni-axial gyroscope that measured the foot instep angular velocity in the sagittal plane. Walking/jogging activities were discriminated by processing gyroscope data from each detected stride. Supervised learning of the classifier was undertaken using reference data from an optical motion analysis system. Remarkably good generalization properties were achieved across tested subjects and gait speeds. Sensitivity and specificity of gait phase detection exceeded 94% and 98%, respectively, with timing errors that were less than 20 ms, on average; the accuracy of walking/jogging discrimination was approximately 99%.