Adaptive sliding mode control strategy based on disturbance observer and neural network for lower limb rehabilitative robot

In this study, to achieve accurate tracking of the desired trajectory during passive control of the lower limb rehabilitation robot, an adaptive sliding mode controller based on disturbance observer and radial basis function neural network (RBFNN) is proposed for the lower limb rehabilitative robot in the presence of uncertain parameters and external bounded disturbances. First, the Euler-Lagrange dynamic model of the lower limb rehabilitative robot is described. Second, a sliding mode controller is designed to stabilize the system with an improved sliding mode reach rate under the assumption that all parameters of the dynamics model are known. To achieve a sliding mode controller without the above assumptions, the proposed adaptive RBFNN and the disturbance observers are employed to compensate for disturbances and the uncertainties in the robot's dynamic mode via feedforward loops. The Lyapunov stability theory is used to prove that the proposed controller has accomplished a significant control effect with excellent performance and the output tracking error can be converted to a very small neighborhood through reasonable design parameters. Finally, the performance of the controller based on the state feedback and state observer are demonstrated by numerical simulations, respectively.