Stiffness adjustment for a single-link robot arm driven by series elastic actuator in muscle training

Series elastic actuators (SEAs) can improve the performance of robots and ensure their safety; therefore, they are widely applied in rehabilitation robots. To effectively exercise muscles in late rehabilitation training, robots utilize different impedance characteristics according to human muscle conditions. In this study, a novel real-time parallel variable-stiffness control method is proposed. The method predicts the muscle stiffness of the human body during the interaction process, and then adjusts the port stiffness of the SEA as needed to improve strength training. The impedance controller is used to meet the port stiffness and stability requirements during interactions. The predictive stiffness control method not only assesses the stiffness of human muscles in real time but also meets the constraint conditions associated with interaction stability, speed, and position. This control structure enables effective and safe strength training. Finally, the experimental results indicate that the stiffness control method based on muscle stiffness prediction can accurately match the port stiffness and provide interaction stability for effective muscle exercise.