Professor David H Owens
Professor of Control and Systems Engineering
Professor D H Owens FREng, BSc, PhD, ARCS, CMath, FIMA, CEng, FIEE, FIMechE
Department of Automatic Control and Systems Engineering
University of Sheffield
Tel: (+44) (0)114 222 5684
Fax: (+44) (0)114 222 5683
Email: d.h.owens @ sheffield.ac.uk
Professor David Owens was the Head of the Department of Automatic Control and Systems Engineering from 1999-2008 and Dean of Engineering at Sheffield University in the period 2002-2006. He was the Convener of the Deans in the period 2005-06. Prior to this he was Head of the School of Engineering and Computer Science at the University of Exeter where he also acted as Chairman of Exeter Enterprises Ltd (the University Technology Transfer Company).
Following 4 years with the UKAEA, he has over 34 years involvement in academic life with experience of departmental teaching, administration and research and high-level University committee service in three UK Universities (Sheffield, Strathclyde and Exeter Universities). He has acted as external examiner for both undergraduate programmes at universities including Glasgow, Bath, Sheffield, Coventry, De Montfort, Salford, Loughborough, KUTKM and UPM-Malaysia.
He has had a strong involvement with the Engineering and related Professions having served on many Committees and the Council of the Institution of Electrical Engineers (IET) and currently serves on its Sector Panel for C&I. He is a member and has chaired the Informatics and Control Group of the Institution of Mechanical Engineers and has been a member of both the IMechE Research and Standards Board and Technology Strategy Committee.
He was the elected Chairman of the United Kingdom Automatic Control Council from 1999-2002 representing the interests of the UK Professional Institutions within IFAC.
He was also elected as a Fellow of the Royal Academy of Engineering in 2008 in recognition of his contributions to control theory and practice and the profession over the past four decades.
In addition he has been elected as a Freeman of the Company of Cutlers in Hallamshire since 2007 and is listed in A&C Black's Who's Who.
He has served on the Health and Safety Commission´s Nuclear Safety Advisory Committee (1995-2006) and has been an independent member of British Energy´s Training Standards and Accreditation Board.
Until 2008, he was an editor of an Institute of Mathematics and Its Applications International Journal, is an Associate Editor of several other journals and regularly contributes to international activities through conference organisation, seminars and invited lectures and committees of the International Federation of Automatic Control (IFAC).
Indexing terms: control systems design; systems modelling; adaptive and robust control; repetitive control; iterative learning systems; nonlinear control systems; control systems applications.
Following his experience of nuclear dynamics and control with the UKAEA (1969-73), his extensive research interests have spanned several areas. He has made substantial contributions in areas including:
- Multivariable control systems design in the frequency domain
- Multivariable root-locus theory
- Approximate modelling for design based on step data
- Repetitive and multi-repetitive control
- Nonlinear, robust and adaptive control
- Multi-dimensional systems theory
- Iterative learning control
- Applied optimization theory
Applications have been in the nuclear, aerospace, automotive, power, manufacturing and process control industries. Over the years, his research has been funded by the Engineering and Physical Sciences Research Council, the European Union and the British Council.
He is the author of over 500 refereed technical publications and the author or co-author of four texts, namely, "Feedback and Multivariable Systems" (Peter, Peregrinus, 1979), "Multivariable and Optimal Systems" (Academic Press, 1981), "Analysis and Control of Multipass Processes" (Research Studies Press, 1982 with J B Edwards) and "Stability Analysis for Linear Repetitive Processes" (Springer-Verlag Lecture Notes in Control and Mathematical Sciences 175, 1992, with E Rogers). A further monograph is expected to be published in 2007.
EPSRC Prof E Rogers (Southampton), D H Owens, "Behavioural models in 2D systems theory", (1995-98) £108,000
EPSRC Technology Foresight Challenge, R A Williams D H Owens et al, "Process Tomography - A New Dimension in Advanced Sensor Technology", (1996-99) £530,401
EU ECSC Project, D H Owens, "Intelligent Control of Boiler Feed Systems" as part of a consortium of EU partners in the area of Systems Operation and Maintenance, (1996-99) Euro 310,000
British Council/DAAD, "Iterative Learning Control for Complex Industrial Processes", (Link to Uni-Hamburg-Harburg University, Germany), (1997-2000) £5,900
EPSRC, E Rogers (Southampton, D H Owens, "Feedback Control of Multi-dimensional Systems", (1998-2001) £118,024
EU TMR Network, D H Owens "Advances in Nonlinear Control", as part of an EU consortium co-ordinated by SUPELEC, Paris, (1998-2002) Euro 129,000
EPSRC, D H Owens, "Multi-periodic Repetitive Control". (2000-2003) £150,000
PPARC, Dr H A Alleyne, D H Owens "A programme of plasma wave and turbulence studies with the CLUSTER II digital wave processor" (2001-2004) £337,797
EPSRC, Professor D H Owens, Professor P J Fleming, Dr S Bennett, Dr N Mort, `Master's Training Package in Control and Systems Engineering´. (2000-2005) £814,944
EPSRC, D H Owens (Sheffield part of a collaboration with Southampton University (P Lewin & E Rogers)). "Iterative Learning Control with Application to Multi-axis Systems" (2002-2005) £201,000
In addition, he has involvement with two EPSRC network grants linking UK academics with academic groups in a number of Japanese universities.
- Joint Editor-in-chief of the IMA International Journal of Mathematical Control and Information
- Member, EPSRC Control and Instrumentation College
- Independent Member of the HSE Nuclear Safety Advisory Committee (NuSAC, 1994-2006) and a Member of its Sub-committee on Research
- Member, British Energy Training Standards and Accreditation Board (2001 - 2006)
- International Federation of Automatic Control (IFAC): Member of Theory Committee, Control Design Methods Committee, Nonlinear Control Committee, Young Researchers Prize Committee.
- Member (1995-2005) and Chair (1999-2002) of the United Kingdom Automatic Control Council (UKACC)
- Institution of Mechanical Engineers: Member of the Mechatronics, Informatics and Control Group Committee
- Institution of Electrical Engineers: Member of IEE Council and the Control and Instrumentation Sector Panel
- Visiting Professor, University of Exeter School of Engineering, Computer Science and Mathematics
Research related activity and recognition has included:
|1||Invited attendee and presenter at the Oberwolfach Conference on Mathematical methods in Control||08/2002|
|2||Invited Lecture to the EPSRC/Bristol workshop on Delay Differential Equations||09/2003|
|3||Plenary lecture on Repetitive and Iterative Control to the International Conference on Systems Engineering, Coventry University||09/2003|
|4||A plenary lecture at the combined IFAC Symposia on "Adaptive and Repetitive Control" and "Periodic Control Systems", Tokyo, August 2004||08/2004|
|5||Postgraduate Lecture course to EPSRC Workshop during UKACC Conference CONTROL 2004||09/2004|
|6||Seminar on ""Iterative Learning Control" at RMIT, Melbourne, Australia||08/2004|
|7||Plenary Lecture at the International symposium NDS,2005 Symposium, Wuppertal||07/2005|
|8||Plenary/invited lectures at the Tokyo Denki Centre of Excellence Workshop on Human Adaptive mechatronics – "Iterative Learning Control"||09/2004
|9||Seminar on Iterative Learning Control at Yokahama University, Japan||03/2006|
|10||International speaker Seminar, EPFL Lausanne "Iterative Learning Control"||11/2005|
|11||Plenary lecture at the Postgraduate Workshop on Advanced Control, Sophia University, Tokyo, Japan||01/2007|
|12||Opening Plenary lecture at the International Conference on Modelling, Identification and Control, Shanghai||07/2007|
|Details of Publication||Brief description of contribution|
|1||OWENS, D H & K FENG (2003), "Parameter Optimisation in Iterative Learning Control", International Journal of Control, 2003, 76(11), pp 1059-1069||Following Owens´ pioneering introduction of the concept of Norm Optimal Iterative Learning Control (ILC) in1996, the paper introduces the paradigm of parameter optimization in ILC. The approach guarantees monotonicity of convergence and provides the first rigorous theoretical convergence analysis. It demonstrates the important link between system positivity and the potential for non-zero limit tracking errors and characterizes the form of the limit set in terms of quadratic equalities. It successfully demonstrates the fact that simple parameterizations are capable of generating excellent performance for practical applications. This paper clearly placed Sheffield as the World leader in this area of study.|
|2||OWENS, D H, LI L & S P BANKS, (2003), Multi-periodic Repetitive Control Systems: a Lyapunov Stability Analysis for MIMO Systems, International Journal of Control, 2003, 77(5), pp 504-515||The paper presents the first rigorous state-space analysis of the stability of linear systems in the presence of multi-periodic demand or disturbance signals and indicates that the results also apply to a class of nonlinear systems. The importance of positivity of the system (or a derived system) is demonstrated and the relevant Lyapunov matrix enables the characterization of both asymptotic stability (asymptotically zero tracking errors) and bounded-input bounded-output stability in a concise manner.|
|3||HATONEN J & D H OWENS, (2004), Convex modifications to an iterative learning control law, Automatica 40, pp 1213-1220||Following Owens´ introduction of the concept of predictive, norm-optimal Iterative Learning Control based on a receding horizon principle in the late 1990s, this paper demonstrates the flexibility implicit within the approach. Convex combinations of current and future iteration input candidates retains the monotonic stability properties and provides a link between performance and robustness properties of the predictive paradigm.|
|4||ROGERS, E. & OWENS, D. H. (2004), On the Stability of Linear Repetitive Processes described by a Delay-Difference Equation. IEEE Transactions on Circuits and Systems I: Fundamental Theory and Applications 51(7):pp. 359-363.||Following the pioneering work of Owens´ and Rogers in the 1980´s and 1990s on the development of effective modelling, analysis and design tools for repetitive/multipass processes, the paper provides a rigorous analysis of the stability of such processes when the model equations contain delays. Delays are a notorious source of control problems and make analysis difficult. This theoretical paper provides a fundamental theoretical base from which further study can proceed.|
|5||Harte T.J. , J Hatonen & D H OWENS (2005), Discrete-time inverse model-based iterative learning control: stability, monotonicity and robustness, Intl Journal of Control, 2005, 78(8), pp 577-586.||The paper provides the first convergence analysis and robustness analysis of the inverse model Iterative Learning Control method. It clearly indicates theimportance of positivity of the multiplicative odelling error (regarded as a matrix between time series) for monotonic mean square convergence and provides a clear link between the dynamic analysis in the state space and frequency response tests. The use of exponential norms enables the analysis of non-monotonic convergence and explains observed phenomena in these terms quite simply.|
|6||OWENS, D H & J J Hätönen (2005), "Iterative Learning Control – An Optimization Paradigm", (Invited paper reviewing the candidate´s contributions in ILC to 2004) IFAC Annual Reviews in Control, 29, 2005, pp 57-70.||Sheffield´s paradigm of parameter optimization in Iterative Learning Control leads the world. This invited paper reviews Sheffield´s contribution to 2004/5 and places it in context for the general reader. It defines the concept and provides a clear presentation of the basic ideas of single and multi-parameter optimal ILC, conditions for convergence and a characterization of the set of limit errors time-series. Gradient and inverse-model-based methods are special cases of this general analysis. The potential benefits of high order ILC parameterizations is defined and conditions provided for the existence of "basis functions" to guarantee convergence to zero error for non-positive plants.|
|7||LIN T, D.H.OWENS, & J HATONEN (2006), "Newton Method based iterative learning control for discrete non-linear systems", International Journal of Control, 2006, 79(10), pp 1263-1276||The paper presents a novel approach to the solution of nonlinear discrete time Iterative Learning Control (ILC) problems using an interpretation of the Newton Method for solving sets of nonlinear algebraic equations. The effect is to decompose the problem into a sequence of linear ILC problems that can be solved by POILC, POILC or other methods. Conditions for semi-local convergence are presented and illustrated using a simple example.|
|8||OWENS D H, M TOMAS RODRIGUEZ & J HATONEN (2006), "Limiting behaviour in Parameter Optimal Iterative Learning Control", International Journal of Automation and Computing, 2006, 3, pp 222-228||The paper presents the first analysis of an observed phenomenon in applications of Parameter Optimal Iterative Learning Control (POILC) where the algorithm suffers from periods of slow convergence followed by accelerated convergence. The paper shows that a defined set of potential limit errors signals can be used to explain the phenomenon by characterizing regions which attract solutions and others which repel. Examples are used to demonstrate the ideas.|
|9||HATONEN J, D H OWENS & K FENG, (2006), "Basis functions and parameter optimization in high-order iterative learning control", Automatica, 42, pp 287-294||High order Iterative Learning Control has received much attention. This paper is the first to demonstrate the applicability of POILC to the high order paradigm and to provide conditions for convergence in terms of positivity of system models G. When positivity is not present, the use of previous iteration input data and/or the addition of fixed "basis functions" is shown to improve convergence. Formally proved theorems demonstrate and characterize the existence of globally convergent POILC algorithms. The number of parameters needed may be large dependent upon the eigenstructure of GT + G but examples show that lower order parameterizations can be effective in practice.|