Instantaneous treadmill speed determination using subject's kinematic data.

The treadmill is frequently used in research and clinical assessments for gait analysis. To evaluate mechanical energy and dynamics of walking, the fluctuations of the treadmill speed have to be taken into account. A new algorithm is presented in this study to determine instantaneous treadmill speed using solely the kinematic data of subjects. The algorithm uses an automatic detection of heel contact (HC) and toe off (TO) during treadmill walking. Kinematic data were collected from two groups (healthy adult, n=11; hemiplegic adult, n=9). The gait events determination is validated by comparison with two visual inspection methods. Our algorithm is able to determine instantaneous treadmill speed with accuracy. In fact, the root mean square error between the computed speed (CS) and the measure speed (MS) was weak with an average value of 0.04+/-0.021ms(-1). So, the computed speed reflects the variations of the belt speed and could be an important contribution to energetic and dynamic gait analysis on a treadmill.