An Improved Generation Scheme of Blended Trajectory for Rehabilitation Robots

Abstract

The trajectory generation ensuring uniform velocity and reduced-jerk is desired for rehabilitation robots. However, this requirement is largely overlooked in the research of rehabilitation robotics. Most research prototypes of existing rehabilitation robots used a cubic polynomial approach in which sudden and large acceleration changes have occurred at the start and the end of the trajectory, which causes theoretically infinite jerk. Moreover, the cubic polynomial approach cannot maintain uniform velocity during the robot's maneuvering. To bridge this gap, in this research, a blended (i.e., hybrid) scheme using fifth-order polynomial at the start and the end of the trajectory and a linear segment in between the polynomials is proposed to generate trajectories for rehabilitation robots. The trajectory generated using the proposed scheme shows no sudden change in acceleration and reduced jerk compared to the cubic polynomial approach.