A New FNN-based Modelling Strategy for Improving Control Capability of Track-Following Performance for Optical Disk Drives Utilizing the Gram-Schmidt Algorithm

W.-N. Huang, C.-C. Teng, W.-P. Chen, and C.-H. Chen (Taiwan)

Keywords

Fuzzy neural network (FNN); GramSchmidt algorithm; optical disk drives (ODDs)

Abstract

A fuzzy neural network (FNN)-based modelling methodology to improve track-following performance (TFP) applying the Gram Schmidt algorithm for optical disk drives (ODDs) is proposed in this paper. Issues related to influence reduction of disturbance effects and parameter variations of system are both addressed and lumped as the assigned functions for verification herein. The tracking performance can be upgraded by the presenting control scheme based on this FNN model, reached to the most remarkable improvement of 78% comparing to the scheme with constant control parameters. The strategy makes use of the projection features by the analysis algorithms, taking the reference command, control input, and output of ODDs into account, to obtain the fastest adjusting mechanism for constructing the effective model in an orthogonal space; meanwhile, the upgrading of performance is shown in distinct comparison results.

Important Links:



Go Back