Power assist control based on learning database of joint angle of powered exoskeleton suitable for wearer’s posture

Abstract

Powered exoskeletons are supposed to be one of the solutions for the shortage of workers caused by declining birthrate and a growing proportion of elderly people. We study a powered exoskeleton that supports worker wearing a radiation protection suit. It is reported that the humidity and temperature in the radiation protection suit are extremely high. EMG sensors are popular to the conventional powered exoskeletons, however, they are not suitable for ours because the high humidity and temperature make the wearer sweat profusely and the sweat make it difficult to use the EMG sensors appropriately. We propose to use 9-axis motion sensors on the wearer instead of EMG sensors to control the powered exoskeleton. The sensors rapidly measure the wearer motion in a high humidity and temperature environment. We evaluate the proposed method using a powered exoskeleton without binding knee in this report.