Characterizing Human Postural Stability by Using Features Extracted From the Correlation Structure of Ground Reaction Force Components

A force plate is one of the most popular tools for postural stability assessment. It decomposes the ground reaction force (GRF) applied to the human body into multiple force components to determine the center of pressure (COP), which is the point at which the GRF is applied. Many COP measures have been proposed to characterize postural stability. Despite postural stability being closely related to GRF components, the interactions between them have rarely been studied. By studying the correlation structure of these GRF components, we developed a set of features to assess postural stability. We determined the correlation matrix of these GRF components and subsequently solved the corresponding eigenvalue problem; we used the resulting eigenvalues to characterize postural stability. The effectiveness of the proposed features was demonstrated by using them to differentiate between individuals in two age groups: 18-24 and 65-73 years. Statistical test results showed that the correlation matrix eigenvalues of the two age groups differed significantly. The classification results demonstrated that most of the correlation matrix eigenvalues were more sensitive to age variations than one of the most reliable and accurate conventional COP features. Furthermore, by reducing the force sensing requirement from three-dimensional to one-dimensional by considering only the vertical GRF components, a simplified version of the proposed approach can be obtained, which could reduce the cost of the force plate system substantially.