报告时间：5月8日（星期一）上午08:30

报告地点：信息楼3楼自动化学院学术报告厅

报告（一）

报 告 人：Krzysztof Galkowski，波兰绿山大学（University of Zielona Gora）教授

报告题目：Iterative learning control, A Linear Repetitive Process Approach

-- Basics and Applications

内容简介：Iterative learning control (ILC) is a technique for controlling systems operating in a repetitive (or pass-to-pass) mode with the requirement that a reference trajectory defined over a finite interval is followed to a high precision. Examples of such systems include robotic manipulators that are required to repeat a given task, chemical batch processes or, more generally, the class of tracking systems. Since the original work of Arimoto in the mid 1980s, the general area of ILC has been the subject of intense research effort. In ILC, a major objective is to achieve convergence of the trial-to-trial error and often this has been treated as the only one that needs to be considered, but then, it is possible that this could lead to unsatisfactory performance along the trial. These can be avoided when noting the clear repetitive (two-dimensional) structure of the ILC process.

This lecture presents basic ideas of ILC based on the repetitive process theory and discuss its applications and future directions. Particularly, an application to the gantry robot is shown and to control of electrical motors (PMSM). Experimental verification for these applications are presented too.

报告人介绍：Dr. Krzysztof Galkowski received his Ph.D. degrees from Technical University of Wroclaw in 1977. After a twenty-year stint at the University of Wroclaw, he joined the University of Zielona Gora in 1996, where he is currently a Professor of Institute of Control and Computation Engineering. Professor Galkowski is an inventor of the effective and still being generalised by other researchers, method of the construction of a state-space realization for the multidimensional (n-variate) transfer function matrices, called Elementary Operation Algorithm. His research interests include multidimensional (nD) systems and repetitive processes-theory and applications, Iterative Learning Control and related numerical methods. He is an author/editor of four monographs/books and over 100 papers in the leading peer reviewed journals and over 180 in the proceedings of international conferences. He has given numerous invited plenary talks for international conferences and in many universities (Europe, USA, Canada, China, Australia, India). He, as well, has prepared numerous special issues for leading journals as IJC, MDSSP and others. He has a strong international co-operation with the Universities of Southampton, UK; the universities of Wuppertal, Rostock and Erlangen-Nurnberg, Germany; University of Hong Kong; University of Poitiers, France; University of Thessaloniki, Greece; East China University of Science and Technology, Shanghai; Harbin Institute of Technology, Harbin; Central South University, Changsha; China University of Geosciences, Wuhan, and many others.

Dr. Krzysztof Galkowski is an associate editor for IET Control Theory and Applications, and a member of editorial board of International Journal of Multidimensional Systems and Signal Processing and International Journal of Control. He served as a member of IPC for several international conferences and co-organised a series of international workshops.

报告（二）

报 告 人：Wojciech Paszke，波兰绿山大学（University of Zielona Gora）副教授

报告题目：A frequency-partitioning approach to design of iterative learning control schemes

内容简介：A broad class of controlled systems, including numerous industrial production systems are of repetitive nature, that is, they have to replicate some operations in consecutive trials, where the main aim is to gain the system response accurately following a desirable reference trajectory. One approach to address this control problem under repetitive errors and disturbances is to apply iterative learning control (ILC) which take a benefit of learning of the feedforward signals for subsequent trials by iteratively updating them based on the control error accumulated in previous inputs. In doing so, high performance can be achieved with a low transient tracking error despite large model uncertainty and some type of disturbances.

Former investigations of ILC show that the combination of a time-wise feedback control scheme and an iteration-wise learning scheme is required and straightforwardly leads to a two-dimensional or repetitive control scheme.

During this presentation, a key benefit of utilization of a two-dimensional/repetitive framework will be emphasized. To design the desired feedback and learning controllers the frequency-domain methods are applied together with robust control theory to formulate the general design framework for the ILC scheme. The resulting problem of determining required feedback and learning controllers is reduced to that of checking the existence of a solution constrained by a set of linear matrix inequalities. That can be efficiently solved in terms of semi-definite programming algorithms, creating a very attractive approach from the point of view of computational burden.

Also, a link between these results and practical requirements for developed control schemes, whichare usually described by multiple frequency domain inequalities in (semi)finite frequency ranges, will be explored. Specifically, with aid of the generalized version of Kalman-Yakubovich-Popov (KYP) lemma, ILC design in finite frequency ranges is performed. This approach allows the designer to specify a frequency range where tracking of the reference signal is required, where, for example, this range could be determined by inspection of frequency spectrums of the signal to be tracked. Therefore, a better tracking performance is expected. Some examples will be also given during this presentation to illustrate the developed results.

报告人介绍：Wojciech Paszke was born in Zielona Gora, Poland, in 1975. He received his M.Sc. and Ph.D. degrees, both in electrical engineering, from the Technical University of Zielona Gora, in 2000 and 2005, respectively. Between 2008 and 2010 he was affiliated to Eindhoven University of Technology, The Netherlands, where he has been a control systems expert on high precision positioning of electron microscope. Currently, he is affiliated to the Institute of Control and System Engineering at the University of Zielona Gora , Poland. His major interest is mainly in the field of multidimensional (nD) systems, repetitive processes, iterative learning control schemes and convex optimization techniques for solving of robust control problems. Wojciech Paszke is serving as Editor for Multidimensional Systems and Signal Processing (Springer-Verlag).