Design and Voluntary Control of Variable Stiffness Exoskeleton Based on sEMG Driven Model

Abstract

Exoskeleton robots are an exciting potential solution for patients with motor dysfunction to restore their daily activities. This letter introduces a variable stiffness exoskeleton robot (VSA-EXO) with variable stiffness actuators and a voluntary control strategy based on sEMG sensing. A sEMG-driven musculoskeletal model estimates joint torque and quasi-stiffness of the human joint through the sEMG signals of some muscles related to the joint. The voluntary control strategy adjusts the degree of assistance according to these detected voluntary efforts of the subjects, and transfers the stiffness adjustment of the human joints to the exoskeleton joints. Unlike the traditional assist-as-needed (ANN), this voluntary control strategy does not need to define a trajectory. Feasibility is demonstrated experimentally using three healthy subjects. Calibration experiments for the musculoskeletal model show that the NRMSE of the estimated torque and the actual torque of the three subjects are lower than 7.21%, proving the effectiveness of the musculoskeletal model. When using the exoskeleton assistance, with an assistance ratio of 0.3, the subjects’ effort is reduced by up to 25.32%.