Adaptive Impedance And Neural Control For Exoskeleton Robot

With the rapid development of control theory and biomedicine,the research of two-arm exoskeleton robot has become a hot topic in the field of science,attracting a large number of scholars.It can not only help the elderly to complete some daily activities,help the workers to complete some heavy load tasks,but also assist hemiplegic patients with hemiplegia for rehabilitation training.However,in the control of two-arm exoskeleton robots,two class of problems are usually met: the instability in the interaction task and the non-linearity caused by input and state constraints.To solve this two problems,in this paper,two corresponding adaptive control methods are presented,and the stability of the controller is analyzed in the framework of Lyapunov stability theory.The main research contents of this paper can be summarized as:An adaptive impedance control method is proposed to solve the instability problem when the two-arm exoskeleton robot is interacting with unknown environment.Based on the observation of the human nervous motion control and the high correlation between human body sEMG signals and joint stiffness,this paper use sEMG signals to obtain the arm endpoint stiffness and a variable stiffness observer is used to compensate for the calculated stiffness.Finally,using the skill transfer to hand the impedance information of human arm to robot,letting robot possess the impedance adjustment mechanism as human hands,ensuring the stability of exoskeleton robots when interacting with unknown environment.Considering the saturated input and state constraints problems existed in the exoskeleton robot,with the uncertainty existing in the model input,this paper proposed a disturbance observer-based adaptive neural control methods,through the use of RBF neural network to approximate the system uncertain items,the disturbance observer compensating for a variety of distractions,such as external disturbance,input nonlinear term and the approximation errors of the system model,etc.,the Barrier Lyapunov Function can keep all the states in their own constraints and the control performance of closed-loop system was guaranteed.