Real-time control strategy for EMG-drive meal assistance robot — my spoon

Abstract

The paper presents an ongoing investigation of applying finite state machines (FSM) to drive a meal assistance robot based on electromyogram (EMG) signals. An adapting single-threshold method on EMG power is proposed to recognize different elevation gestures. Predefined control commands are output by FSM based on the extracted EMG features, and used to operate the robot. The performance of control mode is tested in efficiency and comfortableness by both subjective and objective indices. And the high performance of the EMG-control meal assistance robot makes it feasible for users with upper limbs motor disabilities.