Attractor Divergence as a Metric for Assessing Walking Balance

Abstract

Individuals with Parkinson's disease and the aged have a high prevalence of falls. Since an increase in the number of falls is associated with physical and psychological harm, it is prudent that biomechanical metrics be established that will accurately assess an individual's walking balance. In this investigation, we initially used a simple bipedal walking computer model to theoretically establish the relationship between attractor divergence and walking balance. The Lyapunov exponent was used to quantify the amount of divergence present in the walking attractor. Simulations from our model indicated that attractors that have a greater amount of divergence are more susceptible to falls from external perturbations. Based on the results of our simulations, we conducted an initial experiment to explore if the young, aged and individuals with Parkinson's disease have different degrees of attractor divergence. Our results indicate that individuals with Parkinson's disease and aged have walking patterns with a greater amount of attractor divergence. Based on the results of our simulations, we infer that these participants may have a higher probability of losing their balance.