Comparative Analysis of Sit-to-Stand Motions with Different Movement Strategies Using Gaussian Process Dynamics Model

Abstract

A motion analysis technique utilizing the Gaussian Process Dynamical Model (GPDM) has been proposed to identify characteristic body movement patterns from time-series motion data. This study applies this technique to the analysis of sit-to-stand motion, one of the essential activities of daily living. For stable standing, the body’s center of mass (CoM) must be positioned within the base of support, which is defined by the area of contact between the body and the floor. During the sit-to-stand transition, the CoM undergoes significant displacement, and the lower limbs support the body ’s weight within this base. Based on variations in CoM movement, two primary strategies are recognized: the stability strategy and the momentum strategy. The stability strategy involves leaning the upper body significantly forward so that the CoM enters the base of support before the buttocks lift off the seat. In contrast, the momentum strategy involves accelerating the CoM forward and using this momentum to lift the buttocks before the CoM enters the base of support. Through latent variable analysis using GPDM, it was revealed that the momentum strategy is characterized not only by a shallower forward lean of the trunk compared to the stability strategy, but also by a greater and faster lumbar extension at the moment of buttocks lift-off. These findings demonstrate the potential of GPDM to distinguish subtle kinematic differences between movement strategies.