Modeling Human Upper Limb Trajectories for Reaching Motions on CLEVERarm

Abstract

Given the significant potential for robot-assisted rehabilitation, developing well-planned trajectories plays a crucial role in enhancing the effectiveness of such rehabilitation methods. A critical aspect of this field, particularly concerning the movement of the human upper limb, is the redundancy resolution. In this study, we introduce a novel trajectory planning method aimed at addressing the redundancy resolution in reaching motions related to Activities of Daily Living (ADL). This method is inspired by prior studies on maximum manipulability while emphasizing the natural upper limb posture, particularly the human preference for maintaining a nearly steady elbow position during ADL movements unless, of course, the range of the desired motion requires otherwise. A trajectory-combining approach is developed for generating trajectories in the human configuration space. Additionally, we present a configuration transformation model for human-robot configuration alignment. Experimental results validate the hypothesis of a steady elbow position and combine features from the Minimum Jerk (MJ) and Minimum Angular Jerk (MAJ) methods, demonstrating more natural reaching motions. The configuration transformation model has been successfully tested on the TAMU CLEVERarm, a lightweight and compact upper limb exoskeleton.