Towards developing a human-friendly power assist robot for manipulating heavy objects: special focus on manoeuvrability and object's surface friction

Abstract

A power assist robot system was developed for manipulating objects in cooperation with human. Weight perception was included in robot dynamics and control. The robot was simulated for different conditions. Optimum manoeuvrability conditions for vertical lifting and horizontal manipulation of objects were determined. Psychophysical relationships between actual and perceived weights were determined, and load forces and motion features were analysed for unimanual and bimanual lifting of objects. Then a novel control scheme was implemented that reduced the excessive load forces and accelerations, and thus improved the system performances for unimanual and bimanual lifts. Motions were also analysed for lowering objects with the robot. A feed-forward friction model was introduced that addressed the effects of friction between human’s hand and object’s surfaces on weight perception and load force. The findings can be used to develop human-friendly power assist robots for manipulating heavy objects in industries.