Trajectory Tracking of a Piezoelectric Actuator Using State Compensated Iterative Learning Control

• Main Authors: jhen-cheng wang, 王鎮城
• Other Authors: Fu-Shin Lee
• Format: Others
• Language: en_US
• Published: 2006
• Online Access: http://ndltd.ncl.edu.tw/handle/24056558960523692887

博士 === 華梵大學 === 機電工程研究所 === 94 === This dissertation presents a research on trajectory tracking of a piezoelectric actuator using a state compensated iterative learning control (SCILC) scheme. The SCILC scheme has two main features: a novel design of state compensation for iterative learning control (ILC), and an enhancement of ILC using error filtering. In the SCILC scheme, a state compensation is designed to compensate for the change in states from one iteration to the next. Due to the state compensation, the SCILC scheme can yield more precise tracking control in comparison to conventional ILC schemes. The adoption of error filtering in the SCILC scheme is intended to prevent noise accumulation in iterative learning and ensure long-term stability in executing the scheme. In this dissertation, a fundamental ILC theory is introduced and a general feedback assisted ILC (FB-ILC) scheme is presented. Convergence analysis of the FB-ILC scheme is uniquely performed, and thus convergency of the SCILC scheme is established accordingly. In-depth studies into ILC error filtering are undertaken. Convergence bandwidth of an ILC system is novelly addressed. Essential ILC filters are discussed extensively. Analysis of ILC tracking errors using the Wigner-Ville distribution (WVD) is introduced. A methodology for ILC filter design is proposed by utilizing the WVD and the convergence bandwidth analyses associated with an ILC system. SCILC controllers using essential filters are designed and applied to a tracking task with extensive tracking experiments on a piezoelectric actuator. Results of the extensive experiments confirm the efficacy and robustness of the innovative SCILC scheme in tracking control of the piezoelectric actuator.