Preliminary Design and Control of an Ankle Rehabilitation Robot With Variable Stiffness Actuator

Abstract

Intrinsic robotic compliance is crucial for enhancing the safety of physical human-robot interaction. This study proposes a variable stiffness ankle rehabilitation robot for rehabilitation training. A variable stiffness actuator is designed based on the lever principle, which can synchronously adjust the length of the plate springs force arm using the self-centering mechanism. The cantilever beam flexural deformation theory is adopted to construct an actuator stiffness model. It is solved to the optimal solution using the particle swarm algorithm under the constraints of force and material limits. Based on this, a position and stiffness coordinated control strategy is proposed based on the dynamic model. Finally, simulation experiments are conducted to verify the effectiveness of the variable stiffness model and control strategy. The results demonstrate that the variable stiffness model is highly accurate, and the performance of position tracking control, stiffness cooperative control, and stiffness step control is satisfactory.