Design And Research Of The Multi-joint Soft Lower Limb Exoskeleton And Contrl Methods

Because of its portability and comfort,the soft lower limb exoskeleton has a broad application prospect in the fields of health care,disability assistance,motion protection and enhancement.The soft exoskeleton can effectively reduce the user’s energy consumption by assisting some joint movements,but the more joints there are,the more motors are needed,resulting in the increased weight of the exoskeleton,which impairs the assistance efficiency of the soft exoskeleton.To solve this problem,this paper proposes a multi-joint soft lower limb exoskeleton,and deeply studies the key problems of the exoskeleton robot,such as assistance force curve,soft brace structure,motion intention extraction and control system.Firstly,the motion information of human lower limbs is the basis for the design of multi-joint soft lower limb exoskeleton.By using VICON motion capture system and opensource software OPENSIM co-simulation,kinematics and dynamics parameters of lower limb joints were obtained through inverse kinematics and inverse dynamics,which provided basis for assistance force curve design and motion intention extraction.Secondly,six actuation schemes of single-motor and multi-joint soft exoskeleton were proposed based on time-division multiplexing,and the schemes of plantar flexion and hip flexion assistance were studied in detail,and the corresponding assistance force curve was designed.According to the requirement of dynamically changing the loading path of assistance force,an automatic winding device is designed.At the same time,soft brace was designed based on the principle of maximum stiffness,and the bidirectional actuator and sensing and control system were designed.By combining the characteristics of the bidirectional actuator and the automatic winder,a single motor can assist the movement of multiple joints.In addition,an iterative learning controller based on admittance control was constructed to accurately transfer the assistance force to the target joint.An admittance controller is used to track the target assistance force curve.The P-type iterative learning controller was used to eliminate the periodic error caused by the difference of individual.The variation of Bowden line amount caused by lower limb geometry was eliminated by angle feedforward model.At the same time,the adaptive strategy of wide stride gait is designed.Finally,the actual effect of the multi-joint soft lower limb exoskeleton was evaluated by force tracking experiment and muscle fatigue experiment,and the applicable conditions of the driving scheme were discussed from the quality point of view.