Premium
Nonholonomic system control via particle swarm optimization of a neural‐aided direct gradient descent control
Author(s) -
Duong Sam Chau,
Kinjo Hiroshi,
Uezato Eiho,
Yamamoto Tetsuhiko
Publication year - 2011
Publication title -
ieej transactions on electrical and electronic engineering
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.254
H-Index - 30
eISSN - 1931-4981
pISSN - 1931-4973
DOI - 10.1002/tee.20650
Subject(s) - nonholonomic system , gradient descent , control theory (sociology) , particle swarm optimization , artificial neural network , controller (irrigation) , embedding , computer science , process (computing) , control engineering , control (management) , engineering , mobile robot , artificial intelligence , robot , algorithm , agronomy , biology , operating system
It is well known that the class of nonholonomic systems cannot be asymptotically stabilized by continuous static state feedback controls. It has been reported that the so‐called direct gradient descent control (DGDC) is able to stabilize nonholonomic systems. This article attempts to improve the performance of the DGDC by using neural network (NN) and particle swarm optimization (PSO). A control method is proposed by embedding an NN into the control law derived by the original DGDC. PSO is employed in order to search globally the parameters of the controller without requiring a redesign process of the parameters. To verify the method, the control problems of two typical nonholonomic systems, one being a wheeled mobile robot and the other a rotary crane system, are considered under constraints applied to the systems. Comparative performance tests are carried out, showing that the proposed approach outperforms the original method and a neurocontroller. Also, simulations show that the proposed method is able to control the systems effectively under the given constraints without the need of the redesign process. © 2011 Institute of Electrical Engineers of Japan. Published by John Wiley & Sons, Inc.