Research on Active–Passive Training Control Strategies for Upper Limb Rehabilitation Robot

Due to accidents, upper limb movement disorders have become increasingly common. Training can help restore muscle strength and rebuild neurological function. However, the existing single mode has limitations in adapting to the training needs of different rehabilitation stages. Therefore, this paper conducts research on active–passive training control strategies for an upper limb rehabilitation robot. It establishes an upper limb kinematic model based on the Lagrange method and builds a man–machine integration dynamics model for upper limb rehabilitation in MATLAB (R2016a)/Simulink. A design active controller, passive controller, and switching controller based on PI and feedforward compensation strategies are proposed to improve training control accuracy. The output moment of the system during active training is planned to ensure the safety and stability of the training process. By utilizing neural networks to train sample data during rehabilitation training, the fuzzy rules and membership functions in fuzzy intention recognition algorithm are optimized to improve the accuracy of intention recognition during training. By adopting the independently developed experimental platform for the upper limb rehabilitation robot, active–passive training, intention recognition, and training mode switching are achieved. The results show that the active and passive training processes are smooth, the training intention recognition is accurate, and the switching between active and passive training modes is steady. This verifies the feasibility and effectiveness of the established mathematical model in upper limb rehabilitation training.