Note: This content is accessible to all versions of every browser. However, this browser does not seem to support current Web standards, preventing the display of our site's design details.


HURWITZ Lecture : Constrained Estimation

Classical estimation methods, such as the well known Kalman Filter, deal with unconstrained problems. However, constraints are an integral part of many estimation problems encountered in practice. For example, a well studied problem in digital telecommunications is that of “channel equalization” to compensate intersymbol interference. Here the transmitted signal to be estimated is known to be constrained to a finite set (typically ±1). In this talk we will show how constrained estimation problems of this type can be formulated as receding horizon constrained optimization problems. These problems are similar to the more familiar constrained control problems save that there is an additional degree of freedom in estimation, namely the initial state can also be adjusted. With this caveat, many of the ideas of constrained control can be readily extended to the case of constrained estimation. The talk will illustrate the principles by reference to several practical cases including the problem of channel equalization discussed above. We will conclude the talk by showing that there exists a more technical connection between constrained control and estimation. Indeed, the linear constrained estimation problem turns out to be the formal Lagrangian dual of a particular linear optimal control problem having a special nonquadratic cost. This circle of ideas embellishes known connections between linear quadratic unconstrained control and linear quadratic filtering.

On-demand video
Type of Seminar:
Public Seminar
Prof. Graham Goodwin
Centre for Integrated Dynamics and ControlSchool of Electrical Engineering and Computer ScienceThe University of Newcastle,Callaghan NSW 2308,Australia
Dec 17, 2003   17:15

ETH Zentrum, Gloriastr. 35, Building ETZ E8
Contact Person:

Prof. M. Morari
File Download:

Request a copy of this publication.
Biographical Sketch:
Graham C. Goodwin obtained a B.Sc (Physics), B.E (Electrical Engineering), and Ph.D from the University of New South Wales. From 1970 until 1974 he was a lecturer in the Department of Computing and Control, Imperial College, London. Since 1974 he has been with the Department of Electrical and Computer Engineering, The University of Newcastle, Australia. He is the co-author of seven monographs, four edited volumes, and several hundred technical papers in the areas of filtering, prediction and control. Graham Goodwin is the recipient of many international prizes including the USA Control Systems Society 1999 Hendrik Bode Lecture Prize, a Best Paper award by IEEE Trans. Automatic Control, and Best Engineering Text Book award from the International Federation of Automatic Control. He is currently Professor of Electrical Engineering and Associate Director of the Centre for Integrated Dynamics and Control at the University of Newcastle, Australia. Graham Goodwin is the recipient of an ARC Federation Fellowship; a Fellow of IEEE; an Honorary Fellow of Institute of Engineers, Australia; a Fellow of the Australian Academy of Science; a Fellow of the Australian Academy of Technology, Science and Engineering; a Member of the International Statistical Institute; and a Fellow of the Royal Society, London. He has just been elected to the Swedish Royal Academy of Science, the institution that selects Nobel Prize winners.