Human postural control can be modeled as a complex, non-linear multi-sensory and multi actuation system. Postural control requires a balance between various sensory inputs and the corresponding response in a motor control framework. Various motor control disorders result in a propensity for fall. In this paper, a classification scheme using support vector machines (SVM) is investigated to classify individuals between healthy and those prone to balance disorders including falls. First, a computational model is described to illustrate the proposed approach. It is shown that using SVM, in conjunction with L2 Norm associated with the phase space trajectories, it is possible to classify fallers and non-fallers. This metric can then be utilized for clinical application upon further evaluation. This may benefit clinical evaluation of Parkinson’s subjects by complementing the widely used UPDRS scale.

This content is only available via PDF.
You do not currently have access to this content.