Anthropomorphic design of robotic arm trajectories in assembly cells

Abstract

BACKGROUND: Anthropomorphism is attribution of human form or behavior to non-human agents. Its application in a robot increases occupational safety and user acceptance and reduces the mental effort needed to anticipate robot behavior. OBJECTIVE: The research question focuses on how the anthropomorphic trajectory and velocity profile of a virtual gantry robot affects the predictability of its behavior in a placement task. METHODS: To investigate the research question, we developed a virtual environment consisting of a robotized assembly cell. The robot was given human movements, acquired through the use of an infrared based motion capture system. The experiment compared anthropomorphic and constant velocity profiles. The trajectories were based on human movements of the hand-arm system. The task of the participants was to predict the target position of the placing movement as accurately and quickly as possible. RESULTS: Results show that the anthropomorphic velocity profile leads to a significantly shorter prediction time (α = 0.05). Moreover, the error rate and the mental effort were significantly less for the anthropomorphic velocity profile. Based on these findings, a speed-accuracy trade-off can be excluded. CONCLUSIONS: Participants were able to estimate and predict the target position of the presented movement significantly faster and more accurately when the robot was controlled by the human-like velocity profile.