Title: Study on parameters self-tuning of speed servo system based on LS_SVM model

Authors: Pengzhan Chen, Xiaoqi Tang

Addresses: National NC System Engineering Research Centre, Huazhong University of Science and Technology, Wuhan, 430074, Hubei Province, China. ' National NC System Engineering Research Centre, Huazhong University of Science and Technology, Wuhan, 430074, Hubei Province, China

Abstract: Support vector machine (SVM) can approach any of the non-linear functions, so it can be used to reconstruct the dynamic features of the practical system. Driven by the demand of the parameters self-tuning of the speed servo system, a parameters self-tuning method based on least squares support vector machine (LS_SVM) reference model of the practical system had been put forward. In the parameters tuning process, at first, a variable amplitude triangular wave signal of variable frequency was used to stimulate the system and the input and output data sets under the open-loop state of the actual system were collected; then, by using LS_SVM in learning the data sets, a reference model with similar dynamic characteristics to the actual system was established; finally, a coordinate rotation optimisation algorithm was employed to find the optimum parameters of the practical system based on the achieved LS_SVM reference model, the parameters self-tuning process of speed servo system was completed indirectly. The proposed parameters self-tuning method in this paper was proved effective by simulation results.

Keywords: speed servo systems; system identification; least squares SVM; support vector machine; LS_SVM; self-tuning; coordinate rotatory optimisation; simulation.

DOI: 10.1504/IJMIC.2010.033839

International Journal of Modelling, Identification and Control, 2010 Vol.10 No.1/2, pp.12 - 18

Published online: 02 Jul 2010 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article