Control Of Compliant Actuators Based On Nonlinear Elastic Elements

The early research of upper limb rehabilitation robots is mainly concentrated in developed countries such as Europe,America and japan.At present,the market of medical rehabilitation robots in these countries is increasing year by year.However,investment in upper limb rehabilitation robots in China still exists a big gap with Europe,America and Japan.The research of rehabilitation robots for upper limb physical dysfunction is still a harsh work in China.With population aging in China,upper limb injuries are increased and application value of upper limb rehabilitation robots will be meaningful.Therefore,the domestic development of upper limb rehabilitation robots with core competitiveness has important practical significance.Compared with traditional industrial manipulators,the safety is a primary goal for rehabilitation robots.Therefore,in order to ensure the safety of HRI process,the development of compliant actuator is particularly important.Compared with others domestic compliant actuators in types and performance aspects,an idea for designing compliant actuators is proposed in this research.The relationship can be summarized as "low load,low stiffness;high load,high stiffness".To realize the function of compliant actuators based on matched relationship between stiffness and load,research on nonlinear stiffness compliant actuators(NSCAs)is in mechanical and control level.On mechanical level,NSCAs are developed based on load selection to obtain accurate torque resolution in low load and to acquire high bandwidth in high load on mechanical level,which is tradeoff to realize more compliant and accurate human-robots interaction.On control level,control NSCAs based on its kinematics and dynamics model.Summarizes its dynamics control performance based on simulated and experimented result.There,unfortunately,are still some problems in the accuracy and response speed in position control,so performance of the actuator is necessary to be optimized and improved.Optimization control of NSCA is mainly include two aspects.Firstly,due to difficulty of establishing accurate mathematical model of disturbance and noise,disturbance observer is applied to compensate external disturbance and noises.Then,hysteresis phenomenon of NSCAs can be describe as Bouc-wen hysteresis model.Adopting discretization and parameter identification for a hysteresis model to setup a more applicable algorithm for the nonlinear stiffness of NSCA TO eliminate low precision problem caused by hysteresis.Through the optimization of the above two aspects,NSCA achieves accurate force/torque output interaction performance and laid a good foundation for the development of rehabilitation robot too.