Towards Neuro-Fuzzy Compensated PID Control of Lower Extremity Exoskeleton System for Passive Gait Rehabilitation

The objective of this work is to design a neuro-fuzzy compensated PID control for passive gait rehabilitation using a lower extremity exoskeleton system. A prototype of 6-DOFs exoskeleton device is developed to assist the children of age 8–10 years old. A dummy having a well-matched body attributes to a healthy child (10 years) is utilized in this work to carry out the experimental runs. Kinect-LabVIEW setup is employed to compute the desired joint angles in the sagittal plane for a healthy gait trajectory. The Euler–Lagrange method is utilized to formulate the dynamic analysis of the exoskeleton system. As the performance of existing control strategies for gait rehabilitation devices is still debatable; therefore, a robust neuro-fuzzy compensated PID control strategy is designed in this work. The asymptotic stability of the proposed control scheme is proved mathematically by Lyapunov theorem. Thereafter, the proposed control strategy is implemented on the exoskeleton-dummy setup in real-time and compared with the classical PID control strategy. From experimental runs, the root mean square error (RMSE) for proposed control scheme is found to be less by 40% nearly while tracking the desired gait trajectory. The robustness of the proposed controller is also validated by varying lower limb masses of dummy and by providing an external disturbance. It is observed that proposed controller is more robust to deal with the input disturbance as compared to parametric uncertainty. Finally, the low values of settling time in both the directions ensures the fast convergence of proposed controller.