Human-robot collaboration with supernumerary robotic arm

This project investigates how human arm motion prediction can enable safer and more intuitive collaboration with a supernumerary robotic arm (SRA). Participants will perform reaching and object-manipulation tasks while the system records human trajectories together with robot joint states, end-effector motion, task targets, timestamps, and task outcomes. These synchronized data will be used to predict the intended target and the future trajectory of the human hand, enabling the SRA to anticipate the user’s movement and provide timely assistance. The proposed methods will be evaluated using prediction accuracy, task-completion time, control robustness, user workload, comfort, and safety.