Assistive standing of omni-direetional mobile rehabilitation training robot based on support vector regression algorithm

Abstract

Robot assistive standing-up movement for the elders is one of important ways to solve nursing shortage problem caused by population aging. This paper focuses on standing-up movement using an omni-directional mobile rehabilitation training robot, aims to solve the issue of fast generating aid motion trajectories in different heights and different sitting postures. Control algorithm for predicting assistive standing-up trajectories based on support vector regression machine is presented. Firstly, this article established models of shoulder joints, hip joints, knee joints and ankle joints coordinates, angles and spacing parameters of experimenters in different heights under the process from sitting to standing, and made a parametric description of the models. Then, joint motion trajectories were predicted and compared with actual motion trajectories. In the last, the results show that the regression model predicts joints motion trajectories in the standing-up process accurately, which proves the effectiveness of the proposed algorithm.