Aiming at the problems of delay and limited motion range of traditional hand exoskeletons

a rigid-flexible coupled two-degree-of-freedom rehabilitation exoskeleton controlled by surface electromyography(sEMG)is designed.This design integrates the reliability of sEMG

the high power density characteristic of shape memory alloy(SMA)

and the flexibility of the two-degree-of-freedom structure.Firstly

the generation mechanism of sEMG is analyzed.The signal is preprocessed during collection

and features are obtained through feature extraction methods.Based on the exoskeleton’s structural composition

a mathematical model driven by SMA is constructed.Experimental results show that with the exoskeleton’s assistance

the maximum flexion angle of the metacarpophalangeal joint (MCP) is 30.5°

and the minimum extension angle is 4.1°.For the proximal interphalangeal joint (PIP)

the maximum flexion angle is 40.8°

and the minimum extension angle is 12.5°.The hand’s maximum bearing capacity reaches 4.701N.The decrease in sEMG intensity indicates that the exoskeleton can meet the expected motion requirements.The gesture recognition rate of the convolutional neural network(CNN)reaches 72.85%

providing strong support for the exoskeleton’s accurate control.