Dynamics Modeling and Control Strategy Research on Multi-joint Soft Lower Limb Assistive Exoskeletons

Abstract In response to elderly individuals and stroke patients with weakened lower limb muscle strength, this study proposes a multi-joint soft lower limb assistive exoskeleton designed to assist hip and knee joint flexion and extension during human walking. To achieve smooth control of the flexible exoskeleton, a hardware control system based on STM32 is constructed. The study establishes the theoretical dynamics model and motion characteristic equations for the soft exoskeleton. To minimize assistive errors, a control strategy is designed based on PD iterative learning control methods, and control algorithm simulations are performed. For precise assistance of the soft exoskeleton, a mean prediction method is employed to forecast the gait cycle of the human body. To validate the correctness of the dynamic model and the practicality of the soft exoskeleton, tracking experiments with a dummy and walking assistance experiments with human subjects are conducted separately. The dummy tracking experiment results indicate that the percentage errors of peak assistive values for hip and knee joint flexion are 4.77% and 5.81%, respectively, while for extension, the percentage errors are 8.94% and 9.13%. The human walking assistance experiment results show that the peak assistive values provided by the soft exoskeleton for hip joint flexion and extension reach 152.46N and 150.26N, respectively, while for knee joint flexion and extension, the peak assistive values are 107.64N and 106.13N.