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

Abstract

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.

KEYWORDS:

• Euler-Lagrange
• External disturbance
• Gait rehabilitation
• Lower extremity exoskeleton system
• Neuro-fuzzy compensated PID
• Parametric uncertainty

Acknowledgements

The authors acknowledge the Department of Scientific and Industrial Research, India, for starting the initiative PRISM (Promoting Innovations in Individuals, Start-ups and MSMEs) under which this project is carried out. The authors are also grateful to the amiable support of Mr. Monuranjan Dowarah from Mechatronics and Robotics Laboratory, IIT Guwahati, in performing the research experiments.

Additional information

Notes on contributors

Jyotindra Narayan

Jyotindra Narayan received his Bachelor of Technology (BTech) in mechanical engineering from Uttar Pradesh Technical University, India in 2014. Thereafter, he completed his Master of Engineering (ME) from Thapar University, Patiala in 2017 with the specialization of CAD/CAM and robotics, where he worked on the patient side medical manipulators. Now, he is currently a PhD student in Mechanical Engineering Department at Indian Institute of Technology Guwahati (IIT Guwahati), India. His focused research interests are medical assisted robotics, rehabilitation devices for motion assistance and adaptive as well as intelligent control designs in robotics.

Santosha Kumar Dwivedy

Santosha Kumar Dwivedy received the PhD in mechanical engineering from Indian Institute of Technology Kharagpur (IIT Kharagpur), India in 2000. He is currently professor in Department of Mechanical Engineering at Indian Institute of Technology Guwahati (IIT Guwahati). He was also a visiting professor at Institute of Engineering and Computational Mechanics, University of Stuttgart, Germany under DAAD-IIT faculty exchange scheme. He has over 180 journal and conference publications with a focus on integrating robotics and dynamics in various fields. His research interests include both industrial and medical robotics, biomechanics, nonlinear vibration and control along with the applications. Email: [email protected]