Singular value decomposition-based gait characterization

Abstract

BackgroundHuman gait varies based on personal characteristics, the existence of walking problems, or variability of gait parameters. Identifying the sources of variations is significant in detecting walking problems, designing orthotic/prosthetic products, etc. Research questionsWhat are the types of variations in joint angles and ground reaction forces? How do age, sex, height, weight, and walking speed affect the distribution of the modes? MethodsIn this study, temporal variations in the joint angles and ground reaction forces were obtained using singular value decomposition. Then, the relationships among age, sex, height, weight, walking speed, and the coefficients obtained by singular value decomposition were investigated using Pearson’s correlation coefficient matrix. ResultsThe first mode of joint angles and ground reaction forces represent the overall characteristics; the first six modes of joint angles and the first two modes of ground reaction forces express 99.9% of the gait parameter space. We concluded that the walking speed dramatically affects joint kinematics and ground reaction forces. In addition, the effects of age, gender, height, and weight on the joint kinematics and ground reaction forces were also found, but with less contribution. SignificanceThe temporal behavior of the joint angles and ground reaction forces was expressed using a few coefficients due to singular value decomposition. The effects of age, sex, weight, height, and walking speed on the modes were found. The proposed method can be applied to understand the progression of an abnormality, and design orthotic/prosthetic products etc. in future studies.