Parameters Turning of ADRC based on Neural Network

DOI

10.2991/emcs-16.2016.187How to use a DOI?

Keywords

Tracking differentiator; Gradient information; ADRC controller; Parameters turning; Self-learning

Abstract

Objective:Though ADRC controller shows strong robust character and adaptability, it still exist a lot of shortages, such as the frequency characteristie, stability... ete. Those were not solved in the theories. And the most diffieulty of ADRC's application in the industry is that the parameters are too more and the parameters’adjusting is diffieulty. The engineers go short of the experience of the parameters’ adjusting for the eontroller. To solve the problems that the active disturbance rejection controller has too many parameters and it is very difficult to calculate a set of optimal parameters without determinate turning algorithms.Methods: Neural network can get the gradient information of the controlled object, then use gradient descent method to modify nonlinear combined parameters online, so that the ADRC controller has the ability of self-learning, enhanced adaptive ability of ADRC. Results :The simulation results show that the ADRC controller which has good dynamic and static features, improves the design efficiency. The feasibility and effectiveness of this method is further verified

Copyright

© 2016, the Authors. Published by Atlantis Press.

Open Access

This is an open access article distributed under the CC BY-NC license (http://creativecommons.org/licenses/by-nc/4.0/).

Download article (PDF)

TY  - CONF
AU  - Hongliang Kan
AU  - Fenglong Kan
AU  - Nan Chen
AU  - Qiaoshi Ma
AU  - Xin Wang
AU  - Dongwei Zhang
AU  - Ning Qi
AU  - Bin Wang
PY  - 2016/01
DA  - 2016/01
TI  - Parameters Turning of ADRC based on Neural Network
BT  - Proceedings of the 2016 International Conference on Education, Management, Computer and Society
PB  - Atlantis Press
SP  - 767
EP  - 769
SN  - 2352-538X
UR  - https://doi.org/10.2991/emcs-16.2016.187
DO  - 10.2991/emcs-16.2016.187
ID  - Kan2016/01
ER  -