Gait Planning And Control Of Lower Limb Variable Stiffness Actuator Exoskeleton

Lower-limb exoskeleton can be applied on military,medical and industrial handling and other fields as a strongly human-machine coupling system.Assisted exoskeleton robots have the ability to improve human motion and reduce sports injuries through rigid structural support and actuating equipment.At present,lower-limb exoskeleton is mostly driven by rigid actuator,while the compliant actuator driven exoskeleton has poor load capacity.During human walking,the stiffness of joints and muscles changes in real time,it requires the exoskeleton to have stiffness adjustment ability to adapt to the changes of human stiffness.In addition,the existing exoskeletons mostly use model-based planning method or predefined trajectory planning method,which lack of the ability to adjust to the gait trajectory in real time according to the wearer’s motion status.Therefore,there is a need for a lower-limb exoskeleton with variable physical stiffness and online adjustment capabilities.In view of the above problems,this thesis designs a lower-limb exoskeleton robot driven by rope based on variable stiffness actuator(VSA).An online planning algorithm is designed by combining the Kernelized Movement Primitive(KMP)algorithm with the gait detection algorithm based on neural network model.Firstly,according to the movement rule of human lower limb and the demand of online planning algorithm,a lower limb exoskeleton robot system based on VSA is designed,with hip joint and knee joint as active degrees of freedom.In order to meet the requirements of human-machine coordination,lightweight and flexibility,a ropedriven mechanism and exoskeleton human-machine connection structure are designed.The VSA is designed based on the principle of antagonistic variable stiffness,and a stiffness control model is established for exoskeleton joint position / stiffness control,which realizes the stiffness of VSA from 0.6Nm/degree to complete rigidity..Secondly,the human gait characteristics are analyzed,and the gait detection algorithm is proposed based on the neural network model.The gait data from the standard gait database is collected to establish the Gaussian Mixture Model(GMM).The KMP method is used to plan the gait of the lower-limb exoskeleton robot,and the online planning trajectory based on the gait phase is generated.Through the analysis of the human walking law,the joint stiffness trajectory is established,and the real-time adjustment is carried out through the online planning algorithm to realize the online planning of the gait trajectory and stiffness trajectory of the exoskeleton robot,and the simulation verification is carried out.Then,the joint position / stiffness control of the lower extremity exoskeleton robot is studied.Through the kinematics and dynamics of the exoskeleton robot,combined with the stiffness control model of VSA,the position / stiffness control of the exoskeleton VSA joint is realized,and the gait / stiffness trajectory online planning algorithm is used to form the lower-limb exoskeleton robot control strategy for simulation verification.Finally,the experimental platform of variable stiffness lower limb exoskeleton robot is built.The hardware design and control program design of the exoskeleton prototype are carried out,and the gait detection experiment,joint control experiment,exoskeleton squat assist experiment and online planning experiment are carried out on the platform,in order to verify the efficacy.