Knowledge Acquisition from a Biomechanical System: Human Gait Transition as an Example

Abstract

New approaches that allow a logical link to be established between body parameters and the dynamics of locomotion are attracting increasing interest. We propose a method that obtains knowledge from a biomechanical system. The speed of human gait transition from walking to running was investigated. Employing soft clustering and fuzzy logic principles, we derived the most influential body parameters and logical rules between them which define the preferred transition speed (PTS). The first-order PTS determinants are mass, tibial height and thigh length, while those of the second order are lateral malleolus height and body height. Four logical rules allow PTS values to be predicted with an accuracy of 0.03 m/s when using first-order parameters, and of 0.01 m/s when additionally second-order parameters are included. Compared to previously published studies, these accuracies are the best obtained to date, making our method a promising tool for practical applications.