Enhanced ADRC for sinusoidal trajectory tracking of an upper limb robotic rehabilitation exoskeleton

The Enhanced Active Disturbance Rejection Controller (EADRC) is proposed to perform flexion and extension motions for the shoulder and elbow joints repeatedly and precisely in the sagittal plane using a combination of a nonlinear state error feedback (NLSEF), an extended state observer (ESO), and a finite time stable tracking differentiator (FTSTD). Making use of the Euler-Lagrangian method, the mathematical model of the exoskeleton is derived. A sinusoidal trajectory is supplied as input to a two-link multi-input and multi-output (MIMO) upper limb robotic rehabilitation exoskeleton to perform passive rehabilitation. EADRC tracks this sinusoidal trajectory by estimating the states of a system based on the input-output data of the system and actively removes disturbances. The model parameters are varied by +/- 20% of actual value with constant external disturbance to demonstrate the proposed controller's robustness against uncertainties and disturbances. Stability of the controller is discussed. The proposed strategy was compared with state-of-the-art ESO-based techniques in the simulation. Various performance indicators were used to assess the controller's efficacy which acknowledge that the proposed strategy enhances tracking performance and reliability.