Design and EMG-EEG Fusion-Based Admittance Control of a Hand Exoskeleton With Series Elastic Actuators

This paper proposes an underactuated hand exoskeleton designed to assist in recovering lost grasp function. Structurally, the design incorporates a multi-link coupling mechanism driven by a single motor equipped with a series elastic actuator (SEA). The SEA enables bidirectional compliant drive and fore feedback without the need for a force sensor. The connecting rod is optimized to facilitate the natural flexion and extension of the fingers. For control, an admittance control strategy based on real-time fusion of electromyography (EMG) and electroencephalogram (EEG) signals is proposed. EMG signals are used to estimate muscle strength and control the movement of the exoskeleton. EEG signals reflect the active intention of the subjects, and admittance control adjusts the rehabilitation strategy in real-time. For the first time, the degree of concentration is used as a parameter for subject adjustment of rehabilitation training. Finally, experiments on stiffness calibration, muscle force estimation, and admittance control based on EEG-EMG fusion were conducted. The results indicate that the normalized root-mean-square-error (NRMSE) of stiffness calibration is 8.32%. The average inconsistence of concentration and joint torque (ICJT) is 73.18%. The experimental results indicate that the proposed method can enhance the subjective participation of the subjects in the rehabilitation process, thereby improving the overall rehabilitation outcomes.