In this work, a novel neuro-sliding mode controller (NSMC) is developed for a 7 degree of freedom (DoF) upper limb exoskeleton. Even though the regular sliding mode controller (SMC) is very sufficient tool when the unknown dynamics of the system is time invariant, variation in the unknown dynamics cannot be handled by regular SMC. Therefore, two-layer neural network (NN) is used to approximate the exoskeleton dynamics in the structure of the SMC. Stability of the NSMC is developed by using Lyapunov stability criteria. To validate our theoretical claims and to compare NSMC with regular SMC, simulation results are provided at the end of the paper. In the simulation section, advantage of the NSMC over regular SMC is presented in the presence of time-varying unknown exoskeleton dynamics.
Tubitak
217E138
The work described in this paper is supported by TUBITAK (The Scientific and Technological Research Council of Turkey) under project number 217E138.
217E138
Primary Language | English |
---|---|
Subjects | Electrical Engineering, Mechanical Engineering |
Journal Section | Research Articles |
Authors | |
Project Number | 217E138 |
Publication Date | October 1, 2020 |
Submission Date | March 29, 2020 |
Acceptance Date | August 9, 2020 |
Published in Issue | Year 2020 |
This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License.