Enhanced Generalised Predictive Control with Disturbance Compensation

Generalised Predictive Control (GPC) is an optimal model-based control algorithm that uses an explicit dynamic plant model to predict and optimise the future response of the plant. This thesis investigates the application of GPC in the tracking control of a two-link rigid robotic manipulator. Simula...

Full description

Saved in:

Bibliographic Details
Main Author:	Kwek, Lee Chung
Format:	Thesis
Published:	2014
Subjects:	TJ212-225 Control engineering systems Automatic machinery (General)
Tags:	Add Tag No Tags, Be the first to tag this record!

id	my-mmu-ep.6908
record_format	uketd_dc
spelling	my-mmu-ep.69082017-09-12T16:31:24Z Enhanced Generalised Predictive Control with Disturbance Compensation 2014-10 Kwek, Lee Chung TJ212-225 Control engineering systems. Automatic machinery (General) Generalised Predictive Control (GPC) is an optimal model-based control algorithm that uses an explicit dynamic plant model to predict and optimise the future response of the plant. This thesis investigates the application of GPC in the tracking control of a two-link rigid robotic manipulator. Simulation results reveal that unconstrained GPC attains one order of magnitude improvement in terms of mean square trajectory tracking errors (MSE) as compared to a conventional proportional-derivative (PD) controller. Through constrained optimisation procedure, GPC keeps the control actions within some predeﬁned realisable range while providing satisfactory tracking performance. Motivated by the fact that most industrial robots are required to perform repetitive tasks, the applicability of GPC is extended to a repetitive setting. Exploiting the repetitive nature of the disturbances acting on the robot system, disturbance compensation schemes based on iterative learning control (ILC) are integrated into the GPC framework. Two control strategies, i.e., constrained GPC with learning-based disturbance compensator (CGPC-ILC), and GPC with a robust learning mechanism based on a least mean square error estimator (GPCLMSE), are proposed to provide bounded solutions. While complying with input constraint, CGPC-ILC suﬀers slight performance degradation when the input constraint becomes active. GPC-LMSE demonstrates superior tracking performance while maintaining smooth and bounded control actions. In term of MSE, marked improvement rates of 55.8% with CGPC-ILC and 65.2% with GPC-LMSE over nominal GPC are acquired. 2014-10 Thesis http://shdl.mmu.edu.my/6908/ http://library.mmu.edu.my/diglib/onlinedb/dig_lib.php phd doctoral Multimedia University Faculty of Engineering and Technology
institution	Multimedia University
collection	MMU Institutional Repository
topic	TJ212-225 Control engineering systems Automatic machinery (General)
spellingShingle	TJ212-225 Control engineering systems Automatic machinery (General) Kwek, Lee Chung Enhanced Generalised Predictive Control with Disturbance Compensation
description	Generalised Predictive Control (GPC) is an optimal model-based control algorithm that uses an explicit dynamic plant model to predict and optimise the future response of the plant. This thesis investigates the application of GPC in the tracking control of a two-link rigid robotic manipulator. Simulation results reveal that unconstrained GPC attains one order of magnitude improvement in terms of mean square trajectory tracking errors (MSE) as compared to a conventional proportional-derivative (PD) controller. Through constrained optimisation procedure, GPC keeps the control actions within some predeﬁned realisable range while providing satisfactory tracking performance. Motivated by the fact that most industrial robots are required to perform repetitive tasks, the applicability of GPC is extended to a repetitive setting. Exploiting the repetitive nature of the disturbances acting on the robot system, disturbance compensation schemes based on iterative learning control (ILC) are integrated into the GPC framework. Two control strategies, i.e., constrained GPC with learning-based disturbance compensator (CGPC-ILC), and GPC with a robust learning mechanism based on a least mean square error estimator (GPCLMSE), are proposed to provide bounded solutions. While complying with input constraint, CGPC-ILC suﬀers slight performance degradation when the input constraint becomes active. GPC-LMSE demonstrates superior tracking performance while maintaining smooth and bounded control actions. In term of MSE, marked improvement rates of 55.8% with CGPC-ILC and 65.2% with GPC-LMSE over nominal GPC are acquired.
format	Thesis
qualification_name	Doctor of Philosophy (PhD.)
qualification_level	Doctorate
author	Kwek, Lee Chung
author_facet	Kwek, Lee Chung
author_sort	Kwek, Lee Chung
title	Enhanced Generalised Predictive Control with Disturbance Compensation
title_short	Enhanced Generalised Predictive Control with Disturbance Compensation
title_full	Enhanced Generalised Predictive Control with Disturbance Compensation
title_fullStr	Enhanced Generalised Predictive Control with Disturbance Compensation
title_full_unstemmed	Enhanced Generalised Predictive Control with Disturbance Compensation
title_sort	enhanced generalised predictive control with disturbance compensation
granting_institution	Multimedia University
granting_department	Faculty of Engineering and Technology
publishDate	2014
_version_	1747829646260436992

Enhanced Generalised Predictive Control with Disturbance Compensation

Similar Items