Gait feature with homologous PCA for phase alignment

Abstract

We introduced phase alignment procedure, warping method and proposed homologous analysis for time-series gait data. From this analysis, we revealed the principal features including phase and time variations, while the conventional PCA handled phase variations as amplitude variations. The different gait patterns with extracted features were reconstructed and simulated at heel contact and toe off by FEM. The results showed the difference in plantar pressure and contact area, according to the change of joint angle not to foot shape or property. Our proposed method can analyze the phase and amplitude variations. Moreover, that can reveal the kinetic features for characteristic gait pattern with kinematic features.