Analysis and Causal Formulation Proof of an Optimal Iterative Learning Algorithm英文文献资料.docVIP

Intelligent Control and Automation, 2010, 1, 90-95 doi:10.4236/ica.2010.12010 Published Online November 2010 (http://www.SciRP.org/journal/ica) Analysis and Causal Formulation Proof of an Optimal Iterative Learning Algorithm Vassiliki Vita , Marina Vassilaki , Panagiotis Karampelas , George E. Chatzarakis 1 1 1,2 1 1 Department of Electrical Engineering Educators, School of Pedagogical and Technological Education (ASPETE), Athens, Greece IT Faculty, Hellenic American University, Athens, Greece 2 E-mail: vasvita@yahoo.co.uk, pkarampelas@hau.gr, geaxatz@otenet.gr Received August 12, 2010; revised September 6, 2010; accepted September 19, 2010 Abstract Iterative learning control (ILC) is used to control systems that operate in a repetitive mode, improving track- ing accuracy of the control by transferring data from one repetition of a task, to the next. In this paper an op- timal iterative learning algorithm for discrete linear systems is analyzed and a solution for its attainment is proposed. Finally the mathematical proof of the algorithm’s causal formulation is also provided in its com- plete form, since its implementation requires its causal formulation. Keywords: Casual Formulation, Discrete Systems, Iterative Learning Control 1. Introduction terment process” [3]. This approach is more or less an imitation of the learning process of every intelligent be- ing. Intelligent beings tend to learn by performing a trial (i.e. selecting a control input) and observing what was the end result of this control input selection. After that they try to change their behaviour (i.e. to pick up a new control input) in order to get an improved performance during the next trial. Because, a) the overall idea of ILC and the procedure results in a controller that learn