Energy-Efficient Control Algorithm And Mechanism Design For Rehabilitation Lower Limb Exoskeleton Systems

Currently,rehabilitation lower limb exoskeleton systems have seen increasing clinical applications,yet much of the relevant research has primarily focused on functional realization,while neglecting the energy efficiency of the system.This doctoral dissertation addresses the issue of walking energy efficiency in rehabilitation lower limb exoskeleton systems.It investigates energy-efficient control algorithms and corresponding compliant mechanisms for the ankle joint,knee joint,and the overall exoskeleton system.Subsequently,the eAider rehabilitation lower limb exoskeleton system was developed to validate these algorithms and mechanisms.The primary work and achievements are summarized as follows:To address the issue of high energy consumption during walking caused by the ineffective recovery of body motion energy in the ankle joint of exoskeleton robots during the support phase,this dissertation proposes an energy-efficient control algorithm for exoskeleton systems based on a compliant ankle joint.A mechanical energy loss model for human-exoskeleton system walking is constructed.Based on this model,ankle stiffness of the supporting leg and stride length of the swinging leg during walking are optimized using the gradient descent algorithm,in conjunction with ankle stiffness modeling and parameterized gait models to reduce mechanical energy loss during walking.Simulation results of the above algorithm show that the optimization of ankle stiffness and stride length can reduce the mechanical energy loss of the human-exoskeleton system during walking by 25.1%.To address the issue of high knee joint torque and power output caused by the ground reaction forces at the moment when the swinging leg of exoskeleton robots makes contact with the ground,this dissertation presents an energy-efficient control algorithm for exoskeleton systems based on human-exoskeleton damping matching.Inspired by the mechanism of passive knee flexion in the human body to reduce joint torque,it proposes the use of optimized spring-damping system responses to plan the knee joint trajectory after the swinging leg of the exoskeleton robot touches the ground,thereby reducing torque and power of the knee joint motor.In experiments,the algorithm’s effectiveness is verified by measuring the ground reaction forces.The results indicate that the algorithm can reduce the impact forces applied to the exoskeleton robot’s leg by 22.5%.To address the issue of high energy consumption in exoskeleton systems during walking due to the work done against gravity,this dissertation presents an energy-efficient control algorithm for exoskeleton systems based on compliant walkers.Firstly,a configuration of parallel compliant legs is proposed within the exoskeleton-walker system,allowing the compliant mechanisms of the walker to be parallel to multiple joints of the lower limb exoskeleton robot to enhance the overall walking efficiency.Secondly,based on this configuration,a dynamic model of the human-exoskeleton-walker system is established,and an adaptive control algorithm for support force adjustment,based on extremum-seeking control,is introduced.This leads to real-time optimization of the support force.Simulation and experimental results demonstrate that this algorithm can reduce joint motor energy consumption in the exoskeleton robot by 67.6%under fixed gait parameters,and it can adapt to users with different body weights.To validate the aforementioned algorithms,this dissertation constructed the eAider system and designed and validated a flexibly adjustable stiffness ankle joint along with a flexibly adjustable constant force walker.The above methodology improves the energy efficiency of lower limb exoskeleton systems to a certain extent,and can provide a reference for research in the design and control of flexible mechanisms.