Super Twisting Sliding Mode Controllers And Kalman-Bucy Filter For Single Axis Positioning System

Chiew, Tsung Heng (2018) Super Twisting Sliding Mode Controllers And Kalman-Bucy Filter For Single Axis Positioning System. Doctoral thesis, UTeM.

	Text (24 Pages) Super Twisting Sliding Mode Controllers And Kalman-Bucy Filter For Single Axis Positioning System.pdf - Submitted Version Download (428kB)
	Text (Full Text) Super Twisting Sliding Mode Controllers And Kalman-Bucy Filter For Single Axis Positioning System.pdf - Submitted Version Restricted to Registered users only Download (4MB)

Abstract

Demands for accuracy and precision in machine tools have generated great interests for development of high performance drive control system with excellent characteristics in reference tracking, chattering,and robustness against input disturbance and load variation. Recently, a nonlinear control approach named super twisting sliding mode controller (ST -SMC) becomes attractive for its ability to meet complex demands on system performance where classical controllers have failed to meet.ST-SMC provides good tracking quality and effectively proven disturbance rejection property. However,chattering still exist as an issue in application of ST-SMC. To-date, there exists a knowledge gap in in-depth analyses on optimal design of gains parameters in control laws of ST-SMC constituting trade-off between tracking accuracy and effect of chattering. This thesis presents optimal design of ST-SMC with enhanced smoothening functions for precise tracking performances and reduced chattering; validated on a single axis sliding unit with direct driven linear motor.In addition, a Kalman-Bucy filter (KBF) was designed and applied to estimate velocity signal for the control system thus eliminating effect of noise amplification normally associated with numerical differentiation of position signal. A Taguchi optimization method was applied to optimize the control laws gain parameters of ST-SMC based on three performance indexes, namely; root mean square of tracking error (RMSE), chattering amplitude reduction in frequency domain,and variations in RMSE values from exposure to input disturbance. The optimal values of the gain parameters L and W were 0.7 and 10 10-5 respectively; with a confidence level of 95%.Two variants of ST-SMC were formulated based on modifications of the control laws of original ST-SMC; where the signum function was replaced by either a hyperbolic tangent function or an arc-tangent smoothening function to a form hyperbolic ST-SMC (HST-SMC) and an arc-tangent STSMC (Arc-ST-SMC) respectively. Five controllers were designed and validated experimentally, namely; cascade P/PI controller, pseudo-SMC, optimized ST-SMC, HSTSMC, and Arc-ST-SMC. The control performances of each controller were analyzed with respect to tracking accuracy,chattering suppression, and robustness against input disturbance and system dynamics variation. The optimized ST-SMC produced the best overall control performance with 9.6% (RMSE),3.9% (disturbance rejection),and 13.4% (robustness) superior results compared to the other variants of SMC-based controllers. On the other hand, HST-SMC produced a comparable tracking performance to optimized STSMC with minimal difference of 7.3% (RMSE), 0.4% (disturbance rejection), and 0.7%(robustness).HST-SMC offers a fair trade-off in control performance between tracking accuracy, disturbance rejection and chattering attenuation. Arc-ST-SMC showed its strength with a significant 71.4% reduction in chattering effect.Finally, this thesis has demonstrated outstanding control performances of ST-SMC-based controllers that produced tracking accuracy that was 96.0% better than classical cascade P/PI controller.

Item Type:	Thesis (Doctoral)
Uncontrolled Keywords:	Automatic control, Sliding mode contro, Kalman-bucy filter, Single axis positioning system
Divisions:	Library > Tesis > FKP
Depositing User:	Mohd. Nazir Taib
Date Deposited:	05 Sep 2019 03:00
Last Modified:	17 Mar 2022 15:49
URI:	http://eprints.utem.edu.my/id/eprint/23382
Statistic Details:	View Download Statistic

Actions (login required)

View Item