An Automatic Method to Identify Human Alternative Movement Application to the Ingress Movement

Alternative movement techniques identification is primordial in many studies of biology, medicine and ergonomics. Few studies in literature propose quantitative techniques to achieve this purpose. Park et al. (2005) proposed a biomechanical index based technique named JCV. This method finds its limits when dealing with three dimensional complex movements. In the present study we propose a modification of this method so that it will be applied to complex movements. We consider a non habitual end effector (Midpoint between Hips). We obtain other indices, 3BJCV, to which a non-supervised clustering technique is applied. We have applied our method to the ingress movements of 37 young and elderly subjects with or without prosthesis entering in a minivan vehicle. Our method allows the identification of the two big classes of ingress movements observed by Ait El Menceur et al. (2008b).