2020-10-20T20:24:40Zhttp://edoc.mpg.de/ac_p_oai.ploai:edoc.mpg.de:137352005-11-0719:91
Abstraction based supervisory controller synthesis for high order monotone continuous systems
Moor, T.
Raisch, J.
expertsonly
Abstraction based approaches to hybrid control systems synthesis have so far been mostly limited to problems with low- order linear continuous dynamics. In this paper, results from the theory of monotone dynamical systems are used to efficiently compute discrete abstractions for a class of nonlinear models. Furthermore, a situation is investigated where the high-dimensional plant state converges to a low- dimensional manifold; in the proposed approach the computational effort is governed by the dimension of the low- order manifold without neglecting the high-order dynamics. Results are applied to synthesize a discrete event controller for the automatic start-up of a nonlinear distillation column model of order 42.
Springer-Verlag Berlin
2002
InBook
http://edoc.mpg.de/13735
urn:ISSN:0170-8643
Modelling, analysis, and design of hybrid systems, 247-265 (2002)
en
oai:edoc.mpg.de:137382004-03-1719:91
Discrete supervisory control of hybrid systems based on l- complete approximations
Moor, T.
Raisch, J.
O'Young, S.
expertsonly
hybrid systems; supervisory control; behavioral approach; l- complete approximations
The topic of this paper is the synthesis of discrete supervisory control for hybrid systems Sigma with discrete external signals. Such systems are in general neither l- complete nor can they be represented by finite state machines. Our solution to the control problem is as follows: we find the strongest l-complete approximation (abstraction) Sigma (l) for Sigma, represent it by a finite state machine, and investigate the control problem for the approximation. If a solution exists on the approximation level, we synthesize the maximally permissive supervisor for Sigma (l). We show that it also solves the control problem for the underlying hybrid system Sigma. If no solution exists, approximation accuracy can be increased by computing the strongest k-complete abstraction Sigma (k), k > l. The basic ideas regarding the approximation step are explained within the framework of Willems' behavioral systems theory. Implementation issues are treated in a state space framework, and the main results are interpreted from a traditional control engineering point of view.
2002
Article
http://edoc.mpg.de/13738
urn:ISSN:0924-6703
Discrete Event Dynamic Systems-Theory and Applications, v.12, 83-107 (2002)
en
oai:edoc.mpg.de:138032007-10-2619:91
Population balance modelling and H-infinity-controller design for a crystallization process
Vollmer, U.
Raisch, J.
expertsonly
crystallization; population balance; process control; infinite- dimensional system; robust control
In this paper we consider the robustly stabilizing control of a 10001 draft tube baffled crystallizer. When operated at high fines dissolution rates, the crystallizer exhibits sustained oscillations. A detailed population balance model for the process can be found in the literature. Based on this detailed model we develop a simpler population balance model. This, in turn, permits the derivation of an irrational transfer function from manipulated to measured variable. An H-infinity mixed sensitivity minimization problem is formulated and solved using an infinite-dimensional version of H-infinity theory. Two different controllers are designed and compared in simulation studies.
2002
Article
http://edoc.mpg.de/13803
urn:ISSN:0009-2509
Chemical Engineering Science, v.57, 4401-4414 (2002)
en
oai:edoc.mpg.de:347182005-11-0919:91
Application of virtual reference direct control design in rehabilitation engineering
Previdi, F.
Schauer, T.
Savaresi, S. M.
expertsonly
This paper deals with design of feedback controllers for kneejoint movement of paraplegics using Functional Electrical Stimulation (FES) of the paralysed quadriceps muscle group. The controller design approach, Virtual Reference Feedback Tuning (VRFT), is directly based on open loop measured data and fits the controller in such a way that the closed-loop meets a model reference objective. The use of this strategy, avoiding the modelling step, significantly reduces the time required for controller design and considerably simplifies the rehabilitation protocols. Linear and nonlinear controllers have been designed and experimentally tested, preliminary on a healthy subject and finally on a paraplegic patient. It is shown that the control design is effective in controlling the leg position and rejecting disturbances and it could be successfully used in a rehabilitation protocol.
2002
Conference-Paper
http://edoc.mpg.de/34718
3. Wismarer Automatisierungssymposium (2002)
oai:edoc.mpg.de:347272005-11-0719:91
Controller design for an oscillatory DTB crystallizer based on a population balance model
Vollmer, U.
Raisch, J.
expertsonly
In industrial crystallization, fines dissolution is used to improve crystalsize distribution (CSD) properties. Intense fines dissolution may lead to undesirable oscillations. Therefore, a controller design based on a population balance model is proposed. A late lumping approach is used where the controller design is based on an irrational transfer function resulting from the linearized population balance. The resulting irrational controller transfer function is then approximated by a rational one. The effectiveness of the controller is demonstrated in simulations. The proposed feedback control enables the use of high fines dissolution rates and hence the improvement of CSD properties.
AIDIC
2002
Conference-Paper
http://edoc.mpg.de/34727
urn:ISBN:88-900775-0-6
Chianese, A.: 15th International Symposium on Industrial Crystallization, September, 15 - 18, Sorrento, Italy, AIDIC, 1233-1239 (2002)
oai:edoc.mpg.de:347312005-11-1219:91
Regelung der Trittgeschwindigkeit beim Liegedreiradfahren von Querschnittgelähmten
Schauer, T.
Hunt, K. J.
Negard, N.-O.
Fraser, M. H.
Stewart, W.
expertsonly
Dieser Beitrag befasst sich mit dem Aufbau und der Regelung von Liegedreirädern für querschnittgelähmte Personen. Der Antrieb erfolgt dabei durch elektrische Stimulation der motorisch gelähmten Beinmuskulatur. Kommerziell erhältliche Dreiräder wurden speziell für diesen Zweck instrumentiert, um sie sowohl stationär in Form eines Ergometers als auch mobil für Freizeitzwecke verwenden zu können. Zur optimalen Regelung der Trittgeschwindigkeit wird in diesem Aufsatz ein neuer Ansatz vorgestellt. Ergebnisse mit querschnittgelähmten Probanden werden präsentiert. - This work deals with the design and feedback control of recumbent tricycles for paraplegic subjects propelled by their paralysed legs using electrical stimulation. Commercially available tricycles were instrumented in such a way that the systems can be either used as a stationary ergometer or as mobile device for recreational use. A new method for optimal cadence control is described. Results with paraplegic subjects are given.
2002
Article
http://edoc.mpg.de/34731
at - Automatisierungstechnik, v.50, 271-278 (2002)
oai:edoc.mpg.de:347322005-11-0719:91
Robust nonlinear control of knee-joint angle: A simulation study
Negard, N.-O.
Schauer, T.
Raisch, J.
expertsonly
The aim of this work was to design a robust nonlinear controller for the knee-joint angle by means of Functional Electrical Stimulation (FES). It is well known that a good mechanical model of the body segments can be obtained by first principles and simple pendulum tests, while it is hard to obtain a good of the lectrically elicited muscle contractions for control purposes due to time-variant and nonlinear properties of the muscles combined with limited measurements. For this reason only a linear model with uncertainty bounds is identified. A nonlinear control design approach is performed based on Lyapunov techniques, where uncertainty terms in the model, especially in the actuator dynamics, are dominated with the intention to make the controller robust. This design approach is highly attractive as it can easily be extended to multivariate systems with many body segments and muscle groups, and it allows nonlinear models of the muscle contraction dynamics to be incorporated into the control design. A simulation study of the controller is carried out using a physiologically based model of the muscle contraction dynamics as a substitution of the real plant. Robustness and performance of the designed controller, also with respect to the discrete time implementation and an unmodelled time delay are discussed.
2002
Conference-Paper
http://edoc.mpg.de/34732
3. Wismarer Automatisierungssymposium (2002)
oai:edoc.mpg.de:347332005-11-0919:91
Graybox-modelling of electrically stimulated musculoskeletal systems in paraplegia
Schauer, T.
Previdi, F.
Negard, N.-O.
Raisch, J.
expertsonly
In this paper, a new approach for estimating a nonlinear model of the electrically stimulated quadriceps muscle group under non-isometric conditions is investigated. In order to identify the muscle dynamics (stimulation pulse width-active knee moment relation) from discrete-time angle measurements only, a hybrid model structure is postulated for the shank-quadriceps dynamics. The model consists of a relatively well known time-invariant passive component and an uncertain time-variant active component. Rigid body dynamics, described by the Equation of Motion (EoM), and passive joint properties form the time-variant part. The actuator, i.e. the electrically stimulated muscle group, represents the uncertain time-varying section. A recursive algorithm is outlined for identifying online the stimulated quadriceps muscle group. The algorithm requires EoM and passive joint characteristics to be known a priori. The muscle dynamics represents the product of a continuous-time nonlinear activation dynamics and an nonlinear static contraction function described by a Normalised Radial Basis Function (NRBF) network which has knee-joint angle and angular velocity as input arguments. An Extended Kalman Filter (EKF) approach is chosen to estimate muscle dynamics parameters and to obtain full state estimates of the shank-quadriceps dynamics simultaneously.
2002
Conference-Paper
http://edoc.mpg.de/34733
3. Wismarer Automatisierungssymposium (2002)
oai:edoc.mpg.de:347342005-11-0719:91
Strategic refinements in abstraction based supervisory control of hybrid systems
Moor, T.
Davoren, J. M.
Raisch, J.
expertsonly
2002
Conference-Paper
http://edoc.mpg.de/34734
urn:ISBN:0-7695-1683-1
Silva, M.: Proceedings/ 6th International Workshop on Discrete Event Systems, 329-334 (2002)
oai:edoc.mpg.de:349862005-11-0719:91
Hierarchical control in a behavioural framework
Raisch, J.
Itigin, A.
Moor, T.
expertsonly
This contribution presents some ideas on the synthesis of hierarchical control systems. Hierarchical control is an attempt to handle complex problems by decomposing them into smaller subproblems and reassembling their solutions in a hierarchical structure. A systematic approach to this topic clearly requires a mathematical guarantee that the resulting system of interacting controller units does indeed solve the underlying complex problem. We will argue that behavioural systems theory represents a neat conceptual framework where this can be achieved.
Institute of Control Enginering, Technical University of Szczecin
2002
Conference-Paper
http://edoc.mpg.de/34986
urn:ISBN:88764-66-7
Proceedings MMAR 2002, Institute of Control Enginering, Technical University of Szczecin (2002)
oai:edoc.mpg.de:349942005-11-0719:91
Sliding-mode control of knee-joint angle: experimental results
Schauer, T.
Holderbaum, W.
Hunt, K. J.
expertsonly
The aim of this work was to experimentally validate a sliding-mode controller for the problem of knee-joint angle tracking. The robust controller was designed on the assumption that only a simplified mechanical model is available but not a detailed description of the muscle dynamics. Experiments with an intact subject have been successfully carried out. Performance limitations of this control method due to the effects of sampling and muscle time delay are reported.
Institut Republike Slovenije za rehabilitacijo
2002
Conference-Paper
http://edoc.mpg.de/34994
Karcnik, T.: 7th Annual Conference of the International Functional Electrical StimulationSociety (IFESS2002), Institut Republike Slovenije za rehabilitacijo, 316-318 (2002)
oai:edoc.mpg.de:349962005-11-0719:91
A pilot study of lower-limb FES cycling in paraplegia
Hunt, K. J.
Schauer, T.
Negard, N.-O.
Fraser, M. H.
Stewart, W.
expertsonly
The aim of this work was to develop the engineering methods and apparatus for FES cycling, and to achieve regular periods of mobile FES cycling over significant distances with paraplegic subjects. The study utilised a commercial recumbent tricycle, which was instrumented for stimulation control. Three subjects with a complete spinal cord lesion at level T7-T10 were recruited for the study. After four months of participation in the study the subjects were able to cycle outdoors for distances of up to 3km in a single session. The subjects are also able to cycle continuously and reliably on an indoor cycle trainer for periods of up to 1 h. We conclude that mobile FES cycling over useful distances outdoors is a realistic option for the paraplegic population, even with a low intensity training regime. Future work will involve FES cycling exercise tests to document changes in cardio-pulmonary fitness in SCI subjects.
Institut Republike Slovenije za rehabilitacijo
2002
Conference-Paper
http://edoc.mpg.de/34996
Karcnik, T.: 7th Annual Conference of the International Functional Electrical Stimulation Society (IFESS 2002), Institut Republike Slovenije za rehabilitacijo, 350-352 (2002)
oai:edoc.mpg.de:349972005-11-0719:91
FES-Cycling with electric motor assist
Hunt, K. J.
Stone, B.
Negard, N.-O.
Schauer, T.
Fraser, H. M.
expertsonly
The aim of this study was to develop and test feedback control strategies for integration of electric motor assist and functional electrical stimulation for paraplegic cycling. An integrated two-loop feedback control strategy is proposed, in which the electric motor is driven to provide feedback control of cycling speed, and the cyclist’s leg power is controlled by automatic variation of stimulation intensity. Representative experimental results are presented for one complete-lesion paraplegic subject. The results show accurate feedback control of: cycle cadence; leg power; and external disturbance rejection. We conclude that the addition of electric motor assist, with an integrated feedback control strategy, can compensate for external disturbances, it can greatly increase the overall power output and thereby significantly extend the range of operation, and it can provide optimal operating conditions for the stimulated muscles.
2002
Conference-Paper
http://edoc.mpg.de/34997
1st FESnet Conference, 7-9 (2002)
oai:edoc.mpg.de:351112005-11-0719:91
Discrete approximation and control of high-order nonlinear continuous systems
Moor, T.
Raisch, J.
Itigin, A.
expertsonly
2002
Conference-Paper
http://edoc.mpg.de/35111
Proceedings of the 15th IFAC World Congress on Automatic Control (2002)
oai:edoc.mpg.de:351122003-07-3019:91oai:edoc.mpg.de:351152005-11-0719:91
Think continuous, act discrete: des techniques for continuous systems
Moor, T.
Raisch, J.
expertsonly
Instituto Superior Tecnico
2002
Conference-Paper
http://edoc.mpg.de/35115
Proceedings of the 10th Mediterranean Conference on Control and Automation, Instituto Superior Tecnico (2002)
oai:edoc.mpg.de:351182005-11-0719:91
Discrete models for hybrid systems
Lunze, J.
Raisch, J.
expertsonly
This contribution provides an introduction to the topic of discrete models for hybrid systems. It motivates the use of discrete approximations of hybrid or purely continuous dynamics and comments on two complementary approaches. The article explains the core property of any suitable approximation and describes its major implications.
Springer
2002
InBook
http://edoc.mpg.de/35118
urn:ISBN:3-540-43812-2
Modelling, analysis, and design of hybrid systems, 67-80 (2002)
oai:edoc.mpg.de:624322009-11-1219:91
Η∞-control of a continuous crystallizer
Vollmer, U.
Raisch, J.
expertsonly
Robustly stabilizing control of an open loop oscillatory crystallization process is considered. The crystallizer is described by a popultaion balance model. From this distributed parameter model an irrational transfer function is obtained which has infinitely many poles and thus represents the infinite-dimensional nature of the system. An infinite dimensional Η∞ controller synthesis method is applied to solve the weighted mixed sensitivity problem for this transfer function. This procedure results in an irrational controller. For practical implementation, the controller needs to be approximated by a rational transfer function. The effectiveness of the controller is demonstrated in simulations.
2001
Article
http://edoc.mpg.de/62432
Control Engineering Practice, v.9, 837-845 (2001)
oai:edoc.mpg.de:629712005-11-0719:91
Structure discrimination of continuous models for biochemical reaction networks via finite state machines
Conradi, C.
Stelling, J.
Raisch, J.
expertsonly
2001
Conference-Paper
http://edoc.mpg.de/62971
urn:ISBN:0-7803-6722-7
Proceedings of the 2001 IEEE International Symposium on Intelligent Control ISIC'01, 138-143 (2001)
oai:edoc.mpg.de:629722009-11-1219:91
Computational advantages of a two-level hybrid control architecture
Moor, T.
Raisch, J.
Davoren, J. M.
expertsonly
We investigate a two-level hierarchical architecture for hybrid control. On the top, a discrete supervisory controller acts on quantised measurement information by switching between a finite number of continuous controllers in order to enforce a language inclusion specification. A widely accepted approach to this problem is to first construct a discrete abstraction of the continuous low-level feedback loops and to subsequently resort to DES techniques to solve the high-level synthesis problem. While in principle adopting this approach, we show how to use the structure induced by the low-level controllers to significantly increase computational efficiency of the abstraction procedure. Our methodology enables the system designer to exploit a trade-off between the increase in computational efficiency and the loss in controller flexibility caused by the specific hierarchical structure.
2001
Conference-Paper
http://edoc.mpg.de/62972
urn:ISBN:0-7803-7063-5
Proceedings of the 40th IEEE Conference on Decision and Control, 358-363 (2001)
oai:edoc.mpg.de:629732005-11-0719:91
Modular supervisory control of a class of hybrid systems in a behavioural framework
Moor, T.
Davoren, J. M.
Raisch, J.
expertsonly
2001
Conference-Paper
http://edoc.mpg.de/62973
Proceedings of the European Control Conference ECC2001, 870-875 (2001)
oai:edoc.mpg.de:630622005-11-0719:91
Hierarchical strategies for hybrid process control problems
Raisch, J.
Itigin, A.
Moor, T.
expertsonly
2001
Conference-Paper
http://edoc.mpg.de/63062
Proceedings of the European Control Conference ECC2001, 2534-2539 (2001)
oai:edoc.mpg.de:651562012-12-0619:91
Approximation of multiple switched flow systems for the purpose of control synthesis
Moor, Thomas
Raisch, Jörg
expertsonly
IEEE Customer Service
2000
Conference-Paper
http://edoc.mpg.de/65156
urn:ISBN:0-7803-6638-7
Proceedings of the 39th Conference Decision and Control - Vol. 4, IEEE Customer Service, 3604-3609 (2000)
en
oai:edoc.mpg.de:651672005-11-0719:91
An approach to hybrid control problems based on approximating continuous plant dynamics by nondeterministic automata
Itigin, A.
Raisch, J.
expertsonly
This paper explains how the hybrid control problem of synthesizing a discrete control scheme for a continuous plant model can be solved by approximation continuous plant dynamics by nondeterministic finite automata and then applying standard methods from DES (discrete event systems) theory. A necessary condition is that the behaviour of the approximating automaton model contains the behaviour of the underlying continuous plant model. Then, any controller forcing the approximation to obey the specifications will also force the continuous plant model to satisfy the specifications. Approximating accuracy can be adjusted depending on the degree of complexity of the hybrid control problem to be solved. The method is illustrated by an application example from chemical process control, the synthesis of a discrete feedback strategy for the start-up procedure of a distillation column.
International Institute of Informatics and Systems
2000
Conference-Paper
http://edoc.mpg.de/65167
Sanchez, B.; Hammel, R.; Soriano, M.; Tiako, P.: Proc. SCI2000, International Institute of Informatics and Systems, 217-222 (2000)
oai:edoc.mpg.de:651692005-11-1219:91
Discrete supervisory control of switched linear systems
Franke, D.
Moor, T.
Raisch, J.
expertsonly
This contribution deals with the synthesis of discrete supervisory control for switched linear systems. Via two successive approximation steps, this hybrid control problem is transformed into a purely discrete problem. The latter can then be solved using a slightly modified version of P. J. Ramadge’s and W. M. Wonham’s "supervisory control theory”. By embedding both approximation steps within J. C. Willem’s ”behavioural systems theory”, we can guarantee that any solution of the resulting discrete synthesis problem will also solve the underlying hybrid problem. All algorithms employed within the approximation and the synthesis procedure terminate after a finite number of steps. The approach is illustrated by means of a simple three-plate thermal system.
2000
Article
http://edoc.mpg.de/65169
at - Automatisierungstechnik, v.48, 460-469 (2000)
oai:edoc.mpg.de:652542005-11-0719:91
Hierarchical Control of Hybrid Systems
Raisch, J.
Itigin, A.
Moor, T.
expertsonly
The topic of this contribution is the synthesis of hierarchical control systems for hybrid problems. We outline a method that is based on a hierarchy of models describing the plant at various levels of abstraction and a decomposition of the overall specification. The proposed method captures several key requirements for hierarchical control: it includes the notion of information aggregation between adjacent control levels; it allows for a combination of continuous and discrete event controllers on various levels of the hierarchy; complexity of the synthesis procedure and the resulting control scheme is considerably reduced when compared to an unstructured (and therefore non-hierarchic) approach; finally, it provides a mathematical guarantee that the specified hierarchical interaction between the different controller levels does indeed solve the overall problem. The method is illustrated by an example from process control, the synthesis of a discrete start-up scheme for a distillation column.
Shaker-Verlag
2000
Conference-Paper
http://edoc.mpg.de/65254
Engell, S.: ADPM 2000 Conference Proceedings, Shaker-Verlag, 67-72 (2000)
oai:edoc.mpg.de:652552005-11-0719:91
Complex Systems - Simple Models?
Raisch, J.
expertsonly
In this contribution, the relation between the intended purpose of a model and its required degree of accuracy is investigated. It is based on an intuitive definition of accuracy from J. C. Willems’ ”behavioural systems theory”. We outline a procedure that generates a strictly ordered set of abstractions, or approximations, for a given detailed model, where ordering is in the sense of approximation accuracy. It is particularly useful for the purpose of control synthesis: within the set of abstractions, there exists a unique coarsest, i.e. least accurate, model that allows a given specification to be met. The resulting control scheme is guaranteed to ”work properly” for the underlying detailed model. The approach is illustrated by a number of examples from process control.
Elsevier
2000
Conference-Paper
http://edoc.mpg.de/65255
Biegler, L.T.: Proceedings ADCHEM 2000, Elsevier, 275-286 (2000)
oai:edoc.mpg.de:652562005-11-0719:91
H∞ -control of a continuous crystallizer
Vollmer, U.
Raisch, J.
expertsonly
Robustly stabilizing control of an open loop oscillatory crystallization process is considered. The crystallizer is described by a population balance model. From this distributed parameter model an irrational transfer function is obtained which has infinitely many poles and thus represents the infinite-dimensional nature of the system. An infinite-dimensional H∞ controller synthesis method is applied to solve the weighted mixed sensitivity problem for this transfer function. This procedure results in an irrational controller. For practical implementation, the controller needs to be approximated by a rational transfer function. The effectiveness of the controller is demonstrated in simulations.
2000
Conference-Paper
http://edoc.mpg.de/65256
ADCHEM2000, 117-122 (2000)
oai:edoc.mpg.de:664252005-11-0719:91
Systems and Control Theory at MPI Magdeburg
Raisch, J.
Vollmer, U.
Mayer, E.
Conradi, C.
expertsonly
This contribution provides an overview of various research projects related to systems and control theory at the Max-Planck-Institut für Dynamik komplexer technischer Systeme, Magdeburg. Four application projects are discussed in some detail: feedback control of crystallization processes, automatic start-up of chemical processes, finite automata as models of biological systems, and discrete event methods in transportation systems.
2000
Conference-Paper
http://edoc.mpg.de/66425
Korbicz, J.: Polish-German Symposium on Science, Research and Education, 3-8 (2000)
oai:edoc.mpg.de:664262013-02-0819:91
Discrete abstractions of continuous systems : an input/output point of view
Raisch, J.
expertsonly
This contribution proposes a hierarchy of discrete abstractions for a given continuous model. It adopts an input/output point of view, and starts from the continuous system behaviour B-c (i.e.,the set of all pairs of input and output signals which are compatible with the continuous model equations). The first step is to construct a sequence of behaviours B-l, l=0,1,..., such that B-0 superset of or equal to B-l superset of or equal to ... superset of or equal to B-c. In a second step, non deterministic Moore automata A(l) are generated as minimal realizations for the behaviours B-l. Hence, the continuous base system and its discrete abstractions A(l) form a totally ordered set of models, where ordering is in the sense of set inclusion of model behaviours or, equivalently, in terms of approximation accuracy. Within this set, there exists a uniquely defined "coarsest" (and therefore least complex) model which allows a given set of specifications to been forced by discrete feedback. The ordering property implies that this discrete feedback also forces the continuous base system to obey the specifications.
2000
Article
http://edoc.mpg.de/66426
Mathematical and Computer Modelling of Dynamical Systems, v.6, 6-29 (2000)
oai:edoc.mpg.de:664272005-11-0719:91
Synthesis of hierarchical process control systems based on sequential aggregations
Raisch, J.
Itigin, A.
expertsonly
This contribution outlines a formal synthesis method for hierarchical control systems. It is based on a hierarchy of models describing the plant at various levels of abstraction and a decomposition of the overall specification. The proposed method captures several key requirements: it includes the notion of information aggregation between adjacent control levels; it allows for a combination of continuous and discrete event controllers on various levels of the hierarchy; complexity of the synthesis procedure and the resulting control scheme is considerably reduced when compared to an unstructured (and therefore non-hierarchic) approach; finally, it provides a mathematical guarantee that the specified hierarchical interaction between the different controller levels does indeed solve the overall problem. The method is illustrated by an example from process control, water level regulation in a two-tank laboratory experiment.
ARGESIM
2000
Conference-Paper
http://edoc.mpg.de/66427
urn:ISBN:3-901608-15-X
Troch, I. et al.: Proceedings 3rd Mathmod IMACS Symposium on Mathematical Modelling, ARGESIM, 385-389 (2000)
oai:edoc.mpg.de:664302005-11-0719:91
Automatic generation of switching start-up schemes for chemical processes
Klein, E.
Itigin, A.
Raisch, J.
Kienle, A.
expertsonly
This paper explains how modern techniques from the areas of hybrid and discrete event systems theory can be used to automatize the start-up procedure of chemical processes. As an application example we synthesize a discrete feedback strategy for the automatic start-up of a distillation column.
Pergamon Pr.
2000
Conference-Paper
http://edoc.mpg.de/66430
Pierucci, S.: Escape 10 - 10th European Symposium on Computer Aided Process Engineering, Pergamon Pr., 619-624 (2000)
oai:edoc.mpg.de:1120132005-11-0719:91
Approximating Automata and Discrete Control for Continuous Systems : Two Examples from Process Control
Raisch, J.
Klein, E.
O'Young, S.
Meder, C.
Itigin, A.
expertsonly
A standard problem in hybrid control systems theory is to design discrete, or symbolic, feedback for a given continuous plant. When specifications are discrete, this problem can be solved by first approximating the continuous plant model by a (nondeterministic) automaton, and then synthesizing discrete (supervisory) control for the automaton. A necessary condition is that the approximation behaviour contains the behaviour of the underlying continuous plant model. Then, any controller forcing the approximation to obey the specifications will also force the continuous model to satisfy the specifications. We use a version of this approach which allows adjustment of approximation accuracy to address two simple process control problems: supervisory control of a three tank laboratory experiment and safety enforcement for an evaporator. In both cases, the entire design process is carried through: we first determine a suitable abstraction, compute the minimally restrictive supervisor, and then present examples for closed loop trajectories.
Springer
1999
InBook
http://edoc.mpg.de/112013
urn:ISBN:3-540-65643-X
Hybrid Systems V, 279-303 (1999)
en
oai:edoc.mpg.de:1120142009-11-1219:91
Supervisory control of hybrid systems within a behavioural framework
Moor, T.
Raisch, J.
expertsonly
This contribution addresses the synthesis of supervisory control for hybrid systems Sigma with discrete external signals. Such systems are in general neither l-complete nor can they be represented by finite state machines. We find an l-complete approximation (abstraction) Sigma (l) for Sigma, represent it by a finite state machine, and investigate the control problem for the approximation. If a solution exists, we synthesize the maximally permissive supervisor for Sigma (l). We show that it also solves the control problem for the hybrid system Sigma. If no solution exists, approximation accuracy can be increased by computing a k-complete abstraction Sigma (k), k > l. This paper is entirely set within the framework of Willems’ behavioural systems theory.
1999
Article
http://edoc.mpg.de/112014
Systems & Control Letters, v.38, 157-166 (1999)
en
oai:edoc.mpg.de:1120162005-11-1219:91
Zwei Wege zur Modellierung des qualitativen Verhaltens dynamischer Systeme
Lichtenberg, G.
Lunze, J.
Raisch, J.
expertsonly
In diesem Beitrag werden zwei komplementäre Ansätze zur Bestimmung qualitativer Abstraktionen für gegebene wertekontinuierliche Systeme vorgestellt. Beide Ansätze verwenden nichtdeterministische Automaten als qualitative Approximationen; die Zustandsmengen und die Transitionsstrukturen der Automaten unterscheiden sich aber fundamental: In einem Fall werden sie über die qualitativen Ein- und Ausgangssignale des zu approximierenden Systems definiert, im anderen Fall durch direkte Partitionierung seines Zustandsraumes. Trotz dieser Unterschiede weisen beide Methoden auch eine Reihe wichtiger Gemeinsamkeiten auf: Sie garantieren, dass das Verhalten der diskreten Approximation das Verhalten des kontinuierlichen Systems enthält, und sie erlauben, auf einfache Weise zwischen Genauigkeit und Komplexität der diskreten Approximation abzuwägen.
This contribution deals with discrete, or qualitative, abstractions of continuous dynamic systems. Two complementary approaches are presented. Nondeterministic automata serve as discrete approximations in both cases. State set and transition structure are fundamentally different, however: they are either defined via qualitative input and output signals of the continuous system, or by direct partitioning of the continuous state space. Despite these differences, both methods share a number of important properties: they guarantee that the input/output behaviour of the continuous system is covered by the approximation behaviour, and they allow for trading off approximation accuracy versus complexity.
1999
Article
http://edoc.mpg.de/112016
at-Automatisierungstechnik, v.47, 187-198 (1999)
de
oai:edoc.mpg.de:1167592005-11-0719:91
A hierarchy of discrete abstractions for a hybrid plant
Raisch, J.
expertsonly
This contribution proposes a hierarchy of discrete abstractions for a given hybrid or continuous system with quantized measurements and symbolic control inputs. The continuous (or hybrid) base system and its discrete abstractions form a totally ordered set of models; ordering is in the sense of set inclusion of model behaviours or, equivalently, in terms of approximation accuracy. As, by construction, the behaviour of the base system is a subset of the behaviours of all abstractions in the hierarchy, any of these abstractions can be used for supervisory control synthesis: if a controller forces the behaviour of any abstraction to live within a specified set, it will also make the underlying continuous or hybrid system obey the specifications.
1998
Conference-Paper
http://edoc.mpg.de/116759
Zaytoon, J.: Proc. 3eme Conference Internationale sur l'Automatisation des Processus Mixtes (ADPM'98) : Les Systemes Dynamiques Hybrides, 55-62 (1998)
en
oai:edoc.mpg.de:1167602005-11-0719:91
Estimating reachable states of hybrid systems via l-complete approximations
Moor, T.
Raisch, J.
expertsonly
This contribution treats the estimation of reachable states for timeinvariant hybrid systems. Using the framework provided by Willems behavioural systems theory, we suggest a method based on l-complete approximations, which can be realized by finite statemachines. The approximating behaviour is a superset of the original behaviour. Hence, the estimate of reachable states based on an l-complete approximation can be shown to be conservative, i.e. the exact set of reachable states is guaranteed to be contained in the estimate. Because of this property our method is adequate for verification tasks where the state variable has to remain within a certain specification.
IAAMSAD and the South African Branch of the Academy of Nonlinear Sciences
1998
Conference-Paper
http://edoc.mpg.de/116760
Bajic, V. B.: Proc. International Conference on Systems, Signals, Control, Computers (SSCC'98) : Vol. 3, IAAMSAD and the South African Branch of the Academy of Nonlinear Sciences, 30-34 (1998)
en
oai:edoc.mpg.de:1167612005-11-0719:91
Supervisory control of hybrid systems via l-complete approximations
Moor, T.
Raisch, J.
O'Young, S. D.
expertsonly
This contribution addresses the synthesis of supervisory control for hybrid systems with discrete external signals. Such systems are in general neither l-complete nor can they be represented by finite state machines. We find an l-complete approximation (abstraction) Σl for Σ, represent it by a finite state machine, and investigate the control problem for the approximation. If a solution exists, we synthesize the maximally permissive supervisor for Σl. We show that it also solves the control problem for the hybrid system Σ. If no solution exists, approximation accuracy can be increased by computing a k-complete abstraction Σk, k > l. This paper is entirely set within the framework of Willems' behavioural systems theory.
IEE
1998
Conference-Paper
http://edoc.mpg.de/116761
A. Giua et al.: Proceedings / WODES '98 - International Workshop on Discrete Event Systems, IEE, 426-431 (1998)
en
oai:edoc.mpg.de:1167692005-11-0719:91
Safety enforcement in process control systems : a batch evaporator example
Klein, E.
Raisch, J.
expertsonly
This contribution presents an approximation based safety enforcementscheme for a batch evaporator benchmark problem. Typically, safety enforcement in chemical engineering is a hybrid problem: the plantis (or parts of it are) represented by a continuous model, whereas the safety enforcing device is usually based on discrete intervention rules. Approximating the plant model by an automaton converts this problem into a purely discrete one, provided two conditions hold: the approximation behaviour must contain the continuous model behaviour; then, any control system guaranteeing safe operation for the discrete approximation will also enforce safety for the underlying continuous or hybrid model; the approximation must be "accurate enough" ("its behaviour small enough"), otherwise an appropriate safety enforcement scheme may not exist.
IEE
1998
Conference-Paper
http://edoc.mpg.de/116769
A. Giua et al.: Proc. WODES '98 - International Workshop on Discrete Event Systems, IEE, 327-333 (1998)
en
oai:edoc.mpg.de:1244932009-11-1219:91
Robust nonlinear control of single limb movement
Negard, N.-O.
Schauer, T.
Raisch, J.
expertsonly
The aim of this work was to design a robust nonlinear controller for the kneejoint angle by means of Functional Electrical Stimulation (FES). It is well known that a good mechanical model of the body segments can be obtained by first principles and simple pendulum tests, while it is hard to obtain a good model of the electrically elicited muscle contractions for control purposes due to time-variant and nonlinear properties of the muscles combined with limited measurements. For this reason the muscle activation dynamics are only modeled as a first order linear system. To describe how the muscle contraction depends on joint angle and velocity a simplified linear static function was introduced. Based on this model a nonlinear control design approach is performed applying the back-stepping methodology. This design approach is highly attractive as it can easily be extended to multivariable systems with many body segments and muscle groups, and it allows nonlinear muscle models to be incorporated into the control design. A simulation study of the controller is carried out using a physiologically based model of the neuro-musculoskeletal system as a substitution of the real plant. Robustness and performance of the designed controller, also with respect to the discrete-time implementation and an unmodeled time delay are discussed.
2003
Conference-Paper
http://edoc.mpg.de/124493
Proc. 5th IFAC Symposium on Modelling and Control in Biomedical Systems, 317-322 (2003)
en
oai:edoc.mpg.de:1244962005-11-0719:91
Online identification of the electrically stimulated quadriceps muscle group
Schauer, T.
Previdi, F.
Hunt, K. J.
Negard, N.-O.
Ferchland, E.
Raisch, J.
expertsonly
In this paper, a new approach for estimating a nonlinear model of the electrically stimulated quadriceps muscle group under non-isometric conditions is investigated. In order to identify the muscle dynamics (stimulation pulse with-active knee moment relation) from discrete-time angle measurements only, a hybrid model structure is postulated for the shank-quadriceps dynamics. The model consists of a relatively well known time-invariant passive component and an uncertain time-variant active component. Rigid body dynamics, described by the Equation of Motion (EoM), and passive joint properties form the time-invariant part. The actuator, i.e. the electrically stimulated muscle group, represents the uncertain time-varying section. A recursive algorithm is outlined for identifying online the stimulated quadriceps muscle group. The algorithm requires EoM and passive joint characteristics to be known a priori. The muscle dynamics represent the product of a continuous-time nonlinear activation dynamics and a nonlinear static contraction function described bya Normalised Radial Basis Function (NRBF) network which has knee-joint angle and angular velocity as input arguments. An Extended Kalman Filter (EKF) approach is chosen to estimate muscle dynamics parameters and to obtain full state estimates of the shank quadriceps dynamics simultaneously.
2003
Conference-Paper
http://edoc.mpg.de/124496
Proc. 5th IFAC Symposium on Modelling and Control in Biomedical Systems, 467-472 (2003)
en
oai:edoc.mpg.de:1245242005-11-0719:91
Hybrid control of a cart-pendulum system with restrictions on the travel
Gromov, D.
Raisch, J.
expertsonly
A globally stabilizing controller for a cart-pendulum system is developed. It is shown that the use of hybrid methods can significantly simplify the procedure of control synthesis in the presence of restrictions on the control signal and on the length of travel. A special feature of the proposed approach is that restrictions are embedded into the control algorithm.
2003
Conference-Paper
http://edoc.mpg.de/124524
Proc. PHYSCON 2003 - International Conference Physics and Control, 1231-1235 (2003)
en
oai:edoc.mpg.de:1245252005-11-0719:91
Admissibility criteria for a hierarchical design of hybrid control systems
Moor, T.
Raisch, J.
Davoren, J. M.
expertsonly
Many application relevant hybrid control problems can be decomposed in a hierarchy of control objectives, where each objective refers to a particular time scale and to a particular level of measurement aggregation. Common engineering practice exploits the hierarchical structure to find ad-hoc solutions to hybrid control problems that are far beyond the computational limitations of known methods for the systematic design of hybrid systems. In this paper, we extend a known design method by additional layers between plant and supervisor to represent a control objectives on a hierarchy of time scales and measurement aggregation. We provide a systematic approach that (i) benefits from a wide range of hierarchical decompositions provided by engineering intuition and that (ii) allows for a formal proof that the composition of the individual layers indeed forms an overall solution.
2003
Conference-Paper
http://edoc.mpg.de/124525
Proc. ADHS03 - IFAC, Conference on Analysis and Design of Hybrid Systems, 389-394 (2003)
en
oai:edoc.mpg.de:1245462005-11-0719:91
A two-level hybrid control strategy for the start-up of a coupled distillation plant
Itigin, A.
Raisch, J.
Moor, T.
Kienle, A.
expertsonly
In this contribution, we suggest an efficient hierarchical hybrid control scheme for the start-up of a distillation plant consisting of two coupled distillation columns. The overall control task is decomposed into a set of low-level and high-level subtasks. Each of these subtasks can be solved with much less effort than the original specifications. Low-level tasks are solved by means of continuous feedback, whereas the high-level task is solved by a discrete supervisory controller operating on quantised measurement information and switching between continuous low-level controllers. The approach presented is set within a behavioural framework; it guarantees that the chosen interaction of low-level and high-level control does indeed enforce the specifications.
2003
Conference-Paper
http://edoc.mpg.de/124546
Proc. ECC2003 - European Control Conference 2003 (2003)
en
oai:edoc.mpg.de:1247912009-11-1219:91
Modeling, simulation and stabilizing H∞ control of an oscillating continuous crystallizer with fines dissolution
Motz, S.
Mitrovic, A.
Gilles, E. D.
Vollmer, U.
Raisch, J.
expertsonly
In this contribution, a detailed model for a continuous crystallizer with fines dissolution is derived. The main focus of this article is the identification of physical reasons, responsible for the oscillations of these crystallization plants. In contrast to many other crystallization models used in literature for the investigation of such limit cycles, detailed kinetic expressions for crystal growth and attrition, as well as for the separation of fines in the annular zone, are incorporated. By dynamical simulations of the model and by comparison with measured data, an undesired dissolution of larger crystals can be identified as a possible reason for the appearance of sustained oscillations. Finally, a stabilizing feedback controller is designed using H∞ theory. It is demonstrated in simulations that this controller enables stable operation of the crystallizer even at a high fines dissolution rate.
2003
Article
http://edoc.mpg.de/124791
Chemical Engineering Science, v.58, 3473-3488 (2003)
en
oai:edoc.mpg.de:1247922009-11-1219:91
Control of batch cooling crystallization processes based on orbital flatness
Vollmer, U.
Raisch, J.
expertsonly
In this article it is shown that moment models for batch crystallization processes are orbitally flat. The state dependent time scaling involved in orbital flatness is physically meaningful and leads to a notion of time that is very natural for the crystallization process. A procedure is presented to check if a desired final crystal size distribution (CSD) is achievable and to compute the temporal temperature profile that produces this desired CSD. Furthermore, the problem of dynamic optimization of the crystallizer operation is reformulated based on the system’s flatness property such that the differential equations are eliminated from the optimization problem. In a case study the effectiveness of this optimization is demonstrated.
2003
Article
http://edoc.mpg.de/124792
International Journal of Control, v.76, 1635-1643 (2003)
en
oai:edoc.mpg.de:1250732005-11-0719:91
Time-optimal scheduling for high throughput screening processes using cyclic discrete event models
Mayer, E.
Raisch, J.
expertsonly
A method for solving the scheduling problem for a class of cyclic systems with respect to throughput maximization is presented. Based on discrete events systems (DES) modeling, the scheduling problem can be formulated as a mixed integer linear optimization problem. The method is applied to High Throughput Screening (HTS) problems and illustrated by means of a small example.
ARGESIM, ARGE Simulation News, Univ. of Technology Vienna
2003
Conference-Paper
http://edoc.mpg.de/125073
urn:ISBN:3-901608-24-9
Troch, I.; Breitenecker, F.: Proceedings 4th MATHMOD ; Vol. 2, ARGESIM, ARGE Simulation News, Univ. of Technology Vienna, 1349-1356 (2003)
en
oai:edoc.mpg.de:1736932005-11-0719:91
Throughput-optimal scheduling for cyclically repeated processes
Mayer, E.
Raisch, J.
expertsonly
This contribution deals with cyclically repeated processes where the same job or the same group of jobs (batch) has to be executed a large number of times, each batch following exactly the same time scheme. Each activity of the single batch takes place on one of the plant’s resources, which all have capacity one. It is assumed that a strictly cyclic operation mode is used, i.e. the time offset between the starting times of two consecutive batches is always constant. Normally, batches are nested, i.e. the next batch will start before the previous one is finished. The combination of these requirements constitutes a specific scheduling problem. Based on a discrete event model for cyclically repeated processes, the constraints for this scheduling problem are derived. In order to achive maximium throughput, cycle time is taken as the objective function for the resulting optimization problem. The scheduling problem can then be cast into a mixed integer linear program (MILP). Additionally, some extensions, such as time window constraints, switching times, and buffers with restricted capacity are presented. The method has been applied to high throughput screening (HTS) plants.
2003
Conference-Paper
http://edoc.mpg.de/173693
Proc. MMAR 2003 - 9th IEEE International Conference on Methods and Models in Automation and Robotics, 871-876 (2003)
en
oai:edoc.mpg.de:2004962005-11-0719:91
Discrete control of switched linear systems
Moor, T.
Raisch, J.
expertsonly
Switched linear systems exhibit a continuous state evolving along the continuous flow of time according to linear time in variant differential equations. Furthermore, a discrete interface to the environment is provided, acting on input signals by switching between a finite number of differential equations and generating output signals when the continuous state crosses certain boundaries. We suggest a conservative approximation scheme based on sampling, state partitioning and l-completion realized by a finite past induced state machine. The control problem is investigated on the approximation level. If a solution exists, it also solves the problem for the switched linear system.
1999
Conference-Paper
http://edoc.mpg.de/200496
Proceedings of the 5th European Control Conference ECC'99 (1999)
en
oai:edoc.mpg.de:2076942012-04-1219:91
Control Strategies for Integration of Electric Motor Assist and Functional Electrical Stimulation in Paraplegic Cycling : Utility for Exercise Testing and Mobile Cycling
Hunt, K. J.
Stone, B.
Negard, N.-O.
Schauer, T.
Fraser, M. H.
Cathcart, A. J.
Ferrario, C.
Ward, S. A.
Grant, S.
expertsonly
Aim: The aim of this study was to investigate feedback control strategies for integration of electric motor assist and functional electrical stimulation for paraplegic cycling, with particular focus on development of a testbed for exercise testing in FES cycling, in which both cycling cadence and workrate are simultaneously well controlled and contemporary physiological measures of exercise performance derived. A second aim was to investigate the possible benefits of the approach for mobile, recreational cycling.
Methods: A recumbent tricycle with an auxiliary electric motor is used, which is adapted for paraplegic users, and instrumented for stimulation control. We propose a novel integrated control strategy which simultaneously provides feedback control of leg power output (via automatic adjustment of stimulation intensity) and cycling cadence (via electric motor control). Both loops are designed using system identification and analytical (model-based) feedback design methods. Ventilatory and pulmonary gas exchange response profiles are derived using aportable system for real-time breath-by-breath acquisition.
Results: We provide indicative results from one paraplegic subject in which a series of feedback-control tests illustrate accurate control of cycling cadence, leg power control, and external disturbance rejection. We also provide physiological response profiles from a submaximal exercise step test and a maximal incremental exercise test, as facilitated by the control strategy.
Conclusion: The integrated control strategy is effective in facilitating exercise testing under conditions of well-controlled cadence and power output. Our control approach significantly extends the achievable workrate range and enhances exercise-test sensitivity for FES cycling, thus allowing a more stringent characterisation of physiological response profiles and estimation of key parameters of aerobic function. We further conclude that the control approach can significantly improve the overall performance of mobile recreational cycling.
2004
Article
http://edoc.mpg.de/207694
IEEE Transactions on Neural Systems and Rehabilitation Engineering, v.12, 89-101 (2004)
en
oai:edoc.mpg.de:2076962005-11-0719:91
Data-Driven Control Design for Neuroprotheses : A Virtual Reference Feedback Tuning (VRFT) Approach
Previdi, F.
Schauer, T.
Savaresi, S. M.
Hunt, K. J.
expertsonly
This paper deals with design of feedback controllers for knee joint movement of paraplegics using functional electrical stimulation (FES) of the paralyzed quadriceps muscle group. The controller design approach, virtual reference feedback tuning (VRFT), is directly based on open loop measured data and fits the controller in such a way that the closed-loop meets a model reference objective. The use of this strategy, avoiding the modeling step, significantly reduces the time required for controller design and considerably simplifies the rehabilitation protocols. Linear and nonlinear controllers have been designed and experimentally tested, preliminarily on a healthy subject and finally on a paraplegic patient. Linear controller is effective when applied on small range of knee joint angle. The design of a nonlinear controller allows better performances. It is also shown that the control design is effective in tracking assigned knee angle trajectories and rejecting disturbances.
2004
Article
http://edoc.mpg.de/207696
IEEE Transactions on Control Systems Technology, v.12, 176-182 (2004)
en
oai:edoc.mpg.de:2078692012-04-1119:91
Detection and Filtering of EMG for Assessing Voluntary Muscle Activity during FES
Schauer, T.
Salbert, R. C.
Negard, N.-O.
Raisch, J.
expertsonly
The aim of this work was to develop a laboratory set-up to assess voluntary activity of up to 4 different muscles which undergo simultaneously an artificial electrical stimulation. Voluntary muscle activity is detected from surface EMG. To eliminate stimulation artefacts, the EMG amplifier is automatically muted during stimulation pulses. The measured EMG between stimulation pulses is then considered as superposition of a lower frequency component (m-wave, muscle artefact) caused by electrical stimulation and a higher frequency component caused by voluntary muscle contractions. The latter part is extracted using a digital high pass filter. The entire set-up was realised by using commercially available hardware, which is controlled in soft real-time from a PC running Matlab/SimulinkTM under standard Linux.
2004
Conference-Paper
http://edoc.mpg.de/207869
urn:ISBN:1-85899-191-9
Wood, D.; Taylor, P.: Proceedings of International Functional Electrical Stimulation Society - FESnet 2004, 185-187 (2004)
en
oai:edoc.mpg.de:2079522009-11-1219:91
Time-optimal scheduling for high throughput screening processes using cyclic discrete event models
Mayer, E.
Raisch, J.
expertsonly
A method for solving the scheduling problem for a class of cyclic systems with respect to throughput maximization is presented. Based on discrete events systems (DES) modeling, the scheduling problem can be formulated as a mixed integer linear optimization problem. The method is applied to High Throughput Screening (HTS) problems and illustrated by means of a small example.
2004
Article
http://edoc.mpg.de/207952
Mathematics and Computers in Simulation, v.66, 181-191 (2004)
en
oai:edoc.mpg.de:2079572005-11-0719:91
Optimal hybrid control for switched affine systems under safety and liveness constraints
Corona, D.
Seatzu, C.
Giua, A.
Gromov, D.
Mayer, E.
Raisch, J.
expertsonly
In this contribution, we propose a two-level hierarchical control strategy to solve the problem of minimising a cost function for a switched affine system under safety and liveness constraints. The constraints are handled on the lower level by a discrete supervisory controller designed using l-complete approximation. Its action can be represented as state invariants added to the plant model. In a bottom-up strategy, we can then synthesise a high-level controller, which computes a set of switching tables using the remaining degrees of freedom to optimise a quadratic performance index.
2004
Conference-Paper
http://edoc.mpg.de/207957
CACSD '04 : IEEE International Conference on Computer Aided Control Systems Design, 35-40 (2004)
en
oai:edoc.mpg.de:2079582005-11-0719:91
Modelling and performance evaluation for DES : A Max-Plus Algebra toolbox for Matlab
Stanczyk, J.
Mayer, E.
Raisch, J.
expertsonly
This paper discusses the usefulness of (max; +) algebra as a mathematical modelling framework for discrete event systems (DESs). A Max-Plus Algebra Toolbox developed at Lehrstuhl für Systemtheorie technischer Prozesse is presented. This software package is a set of functions to take advantage of the (max; +) algebra in the Matlab environment for rapid prototyping, design, and analysis of DESs. An overview of the modelling and analysis concepts of the (max; +) algebra approach for DES is given. Application examples are provided in the final part of the paper to illustrate the potential of this approach and the toolbox.
2004
Conference-Paper
http://edoc.mpg.de/207958
ICINCO 2004 : International Conference on Informatics in Control, Automation and Robotics 2004, 270-275 (2004)
en
oai:edoc.mpg.de:2079592005-11-0719:91
A novel hierarchical control architecture for a class of discrete-event systems
Li, D.
Mayer, E.
Raisch, J.
expertsonly
This paper proposes a novel hierarchical control architecture for a class of discrete-event systems. Under the proposed control scheme, a max-plus algebra model is introduced on the upper level to provide an optimal online plan. On the lower (implementation) level, min-plus algebra is used to solve cooperation problems between sub-plants as well as problems caused by unexpected events. A simple rail traffic example is given to show the effectiveness of the idea.
2004
Conference-Paper
http://edoc.mpg.de/207959
Proc. WODES 2004 7th IFAC Workshop on Discrete Event Systems, 415-420 (2004)
en
oai:edoc.mpg.de:2080712005-11-0719:91
Feedforward control of batch crystallizers : an approach based on orbital flatness
Vollmer, U.
Raisch, J.
expertsonly
This contribution addresses the problem of feedforward control design for batch crystallisers based on moment models. Moment models, which are standard in the crystallisation literature, are shown to be orbitally flat, i.e. they are flat after an appropriate time scaling. The reciprocal of the crystal growth rate serves as the time scaling function such that a new notion of 'time' is defined by the increase in crystal length. For any desired final crystal size distribution (CSD) which is compatible with the crystallisier model it is possible, exploiting flatness, to analytically compute the corresponding temperature trajectory.
2004
Conference-Paper
http://edoc.mpg.de/208071
Allgöwer, Frank; Gao, Furong: ADCHEM 2003 : 7th International Symposium on Advanced Control of Chemical Processes - Preprints; Vol. 1, 489-494 (2004)
en
oai:edoc.mpg.de:2082392005-11-0719:91
Modelling and optimization for high-through-put-screening systems
Mayer, E.
Raisch, J.
expertsonly
The problem of cyclic scheduling under the requirement of throughput maximization is considered for a special class of cyclically repeated batch processes. All batches have to follow an identical time scheme. The same resource may be visited more than once by the same batch and time window constraints may be stated by the user. It is shown that the cyclic scheduling problem can be transformed into a mixed integer linear optimization problem. The method's application to High-Throughput-Screening processes is demonstrated.
2004
Conference-Paper
http://edoc.mpg.de/208239
Allgöwer, Frank; Gao, Furong: ADCHEM 2003 : 7th International Symposium on Advanced Control of Chemical Processes - Preprints; Vol. 1, 513-518 (2004)
en
oai:edoc.mpg.de:2082752012-03-0119:91
Step length estimation of gait by means of inertial sensors
Negard, N.-O.
Schauer, T.
Raisch, J.
expertsonly
Im Rahmen dieser Arbeit ist die Möglichkeit des Einsatzes von Inertialsensoren für die Regelung der Funktionellen Elektrischen Stimulation (FES) beim Gangtraining von Schlaganfallpatienten untersucht worden. Mittels der Miniatur-Inertialsensoren werden die Schrittlänge und -höhe während des Ganges bestimmt. Diese Größen sollen später durch gezielte Anpassung der Elektrostimulation beeinflusst werden. Die verwendeten kommerziellen Inertialsensoren bestehen aus 3 Akzelerometer und 3 Gyroskopen welche in 3 zueinander orthogonalen Richtungen montiert sind. Zur Schätzung der Sensorlage wurde ein indirekter Kalman-Filter entworfen, welcher die Orientierung des Sensors in Bezug auf ein globales Koordinatensystem schätzt. Unter der Annahme, dass keine Kräfte auf den Sensor wirken, werden neben den Winkelraten die Beschleunigungsmessungen zur Berechnung der Sensorlage verwendet. Unter Zuhilfenahme der geschätzten Sensorlage können die Raumkurve des Sensors und somit die Schrittlänge und -höhe offline durch Doppelintegration der Beschleunigungen in einem globalen Koordinatensystem ermittelt werden. Konstante systematische Messfehler werden über Randbedingungen bei der Integration ermittelt und eliminiert. Somit kann eine starke Drift in der Positionsschätzung vermieden werden.
2005
Conference-Paper
http://edoc.mpg.de/208275
urn:ISBN:3-00-013509-X
Automed 5. Workshop 2004, 59-60 (2005)
de
oai:edoc.mpg.de:2082762012-03-0119:91
Geregelte Elektrostimulationsergometrie bei Schlaganfall
Salbert, R. C.
Schauer, T.
Negard, N.-O.
Raisch, J.
Schönle, P. W.
Engler, V.
expertsonly
Für eine bessere Rehabilitation von hemiparetischen Patienten (z. B. nach einem Schlaganfall) wurde ein kommerzielles Fahrradergometer mit Hilfsmotor so weiterentwickelt, dass das motorische Defizit der gelähmten unteren Extremität beim Treten durch elektrische Stimulation der Muskulatur kompensiert werden kann. Im Gegensatz zum herkömmlichen Fahrradfahren von komplett querschnittgelähmten Patienten mittels elektrischer Stimulation werden an die Regelung der Stimulationsintensitäten höhere Ansprüche gestellt. Die Interaktion von Fahrer und Ergometer über das nicht betroffende Bein ist bei Anpassung der Stimulationsintensitäten entsprechend zu berücksichtigen. Die Erfassung der vom gesunden Bein eingebrachten "Antriebsleistung" und deren Verwendung als Vorgabe für das gelähmte Bein stellt einen möglichen Ansatz dar. Zur Realisierung des gewünschten Drehmoments durch das betroffene Bein wurde eine iterative Regelung entwickelt, welche die Stimulationsintensität von Tretzyklus zu Tretzyklus gezielt verändert. Die Aktivitäten der Beine wurden an Hand des Motorstromes bewertet. Erste experimentelle Untersuchungen mit dem entwickelten System werden vorgestellt.
2005
Conference-Paper
http://edoc.mpg.de/208276
urn:ISBN:3-00-013509-X
Automed 5. Workshop 2004, 93-94 (2005)
de
oai:edoc.mpg.de:2372782012-04-1419:91
Regelung linearer parameterveränderlicher (LPV-) Systeme mit Hilfe klassischer Regelungsstrukturen und Anwendung auf nichtlineare Regelstrecken
Sommer, S.
expertsonly
2004
Article
http://edoc.mpg.de/237278
At - Automatisierungstechnik, v.52, 397-398 (2004)
de
oai:edoc.mpg.de:2375522005-11-1519:91
Using chemical reaction network theory to discard a kinetic mechanism hypothesis
Conradi, C.
Saez-Rodriguez, J.
Gilles, E. D.
Raisch, J.
expertsonly
Feinberg s Chemical Reaction Network Theory (CRNT) connects the structure of a biochemical reaction network to qualitative properties of the corresponding system of ordinary differential equations. In particular, no information about parameter values is needed. As such, it seems to be well suited for application in systems biology, where parameter uncertainty is predominant. However, its application in this area is rare, at best. To demonstrate potential benefits from the application, different reaction networks representing a single layer of the well studied Mitogen-activated protein (MAPK) cascade are analyzed. Recent results from Markevich et al. (2004) show that, unexpectedly, multilayered protein kinase cascades can exhibit multistationarity even on a single cascade level. Using CRNT, we show that their assumption of a distributive mechanism for double phosphorylation and dephosphorylation is crucial for multistationarity on the single cascade level.
2005
Conference-Paper
http://edoc.mpg.de/237552
FOSBE 2005 : Foundations of Systems Biology in Engineering : Plenary & Contributed Papers, 325-328 (2005)
en
oai:edoc.mpg.de:2376332005-11-0319:91
A new hierarchical control scheme for a class of cyclically repeated discrete-event systems
Li, D.
Mayer, E.
Raisch, J.
expertsonly
We extend the hierarchical control method in (Li et al., 2004) to a more generic setting which involves cyclically repeated processes. A hierarchical architecture is presented to facilitate control synthesis. Specifically, a conservative max-plus model for cyclically repeated processes is introduced on the upper level which provides an optimal online plan list. An enhanced min-plus algebra based scheme on the lower level not only handles unexpected events but, more importantly, addresses cooperation issues between sub-plants and different cycles. A rail traffic example is given to demonstrate the effectiveness of the proposed approach.
2005
Conference-Paper
http://edoc.mpg.de/237633
Proceedings of the Second International conference on informatics in control, automation and robotics (ICINCO 2005) ; Signal Processing, Systems Modeling and Control : Vol. IV, 30-36 (2005)
en
oai:edoc.mpg.de:2376352013-02-0819:91
Online identification and nonlinear control of the electrically stimulated quadriceps muscle
Schauer, T.
Negard, N.-O.
Previdi, F.
Hunt, K. J.
Fraser, M. H.
Ferchland, E.
Raisch, J.
expertsonly
electrical stimulation; extended Kalman filter; physiological model; neural network; nonlinear control
A new approach for estimating nonlinear models of the electrically stimulated quadriceps muscle group under nonisometric conditions is investigated. The model can be used for designing controlled neuro-prostheses. In order to identify the muscle dynamics (stimulation pulsewidth-active knee moment relation) from discrete-time angle measurements only, a hybrid model structure is postulated for the shank-quadriceps dynamics. The model consists of a relatively well known time-invariant passive component and an uncertain time-variant active component. Rigid body dynamics, described by the Equation of Motion (EoM), and passive joint properties form the time-invariant part. The actuator, i.e. the electrically stimulated muscle group, represents the uncertain time-varying section. A recursive algorithm is outlined for identifying online the stimulated quadriceps muscle group. The algorithm requires EoM and passive joint characteristics to be known a priori. The muscle dynamics represent the product of a continuous-time nonlinear activation dynamics and a nonlinear static contraction function described by a Normalised Radial Basis Function (NRBF) network which has knee-joint angle and angular velocity as input arguments. An Extended Kalman Filter (EKF) approach is chosen to estimate muscle dynamics parameters and to obtain full state estimates of the shank-quadriceps dynamics simultaneously. The latter is important for implementing state feedback controllers. A nonlinear state feedback controller using the backstepping method is explicitly designed whereas the model was identified a priori using the developed identification procedure.
Copyright © 2004 Elsevier Ltd. All rights reserved. [accessed February 8th 2013]
2005
Article
http://edoc.mpg.de/237635
Control Engineering Practice, v.13, 1207-1219 (2005)
en
oai:edoc.mpg.de:2376362005-11-0719:91
Analysis of different crystalliser configurations to perform preferential crystallisation
Elsner, M. P.
Alonso Muslera, E.
Angelov, I.
Fernandez Menendez, D.
Lorenz, H.
Polenske, D.
Vollmer, U.
Raisch, J.
Seidel-Morgenstern, A.
expertsonly
The intention of this paper is the analysis of general concepts for preferential crystallisation with the focus on aspects of quantification. In this preliminary examination, the amino acid threonine was used as a model system. The use of on-line polarimetry in combination with on-line densimetry as well as microscopic investigation enables to obtain some indispensable information about the crystallisation kinetics. Taking these estimated values into consideration a simplified mathematical description as a first approach including a population balance model was established for the simulation of the time changes of liquid phase composition during the preferential crystallisation process. Based on this simplified model different crystalliser configurations were investtigated.
2005
Conference-Paper
http://edoc.mpg.de/237636
16th International Symposium on Industrial Crystallization, 829-835 (2005)
en
oai:edoc.mpg.de:2376442006-05-1819:91
Application Programming Interface and PC control for the 8 channel stimulator MOTIONSTIM8
Negard, N.-O.
Schauer, T.
de Gersigny, J.
Hesse, S.
Raisch, J.
expertsonly
This work describes an application programming interface (API) for the commercially available 8 channel stimulator MOTIONSTIM8. The stimulator for transcutaneous FES applications is manufactured by the company Medel GmbH (Hamburg, Germany) and certified for clinical use. By use of the API, customised stimulation programs can be developed, including closed-loop applications. The user can therefore access various hardware interfaces. The functionality of the API will be outlined and an example application to an electro-mechanical gait trainer is described. For research purposes, computer control of the stimulator was further realised whereas the stimulator can be externally controlled through a serial interface (RS232).
2005
Conference-Paper
http://edoc.mpg.de/237644
The 10th Annual Conference of the International Functional Electrical Stimulation Society : IFESS 2005 : Conference Proceedings, 27-29 (2005)
en
oai:edoc.mpg.de:2378232012-12-0619:91
Optimal Control of Discrete-time Hybrid Automata under Safety and Liveness Constraints
Gromov, D.
Mayer, E.
Raisch, J.
Corona, D.
Seatzu, C.
Giua, A.
expertsonly
2005
Conference-Paper
http://edoc.mpg.de/237823
urn:ISBN:0-7803-8937-9
Proceedings of the 2005 IEEE International Symposium on Intelligent Control - Vol. 1, 243-249 (2005)
en
oai:edoc.mpg.de:2378242009-11-1219:91
Optimization of initial conditions for preferential crystallization
Angelov, I.
Raisch, J.
Elsner, M. P.
Seidel-Morgenstern, A.
expertsonly
The presented paper addresses the optimization of isothermal preferential crystallization for separation of pure enantiomer from racemic mixtures. The optimization uses a rigorous process model. A sequence of stochastic and deterministic minimization algorithms are used. Optimized are the initial values of mass of racemate, mass of seeds and mean seed size. The approach is illustrated by a case study for amino acid L-/D-threonine. A promising cyclic operation mode is discussed. Although the results are not validated experimentally, the aim of this work is increase the applicability of the process.
2006
Article
http://edoc.mpg.de/237824
Industrial & Engineering Chemistry Research, v.45, 759-766 (2006)
en
oai:edoc.mpg.de:2378282005-11-1719:91
Optimal control strategies for batch-wise operation of preferential crystallization for gaining high-purity enantiomers
Angelov, I.
Raisch, J.
Elsner, M. P.
Seidel-Morgenstern, A.
expertsonly
Preferential crystallization is an effective and comparatively cheap technology for enantioseparation at different scales. It is based on different crystallization rates of the two enantiomers in the presence of seed crystals of a single optically active isomer and thus allows the selective crystallization of just one enantiomer in a certain time period. The work to be presented here focuses on new concepts for performing preferential crystallization separations. The application of online polarimetry combined with refractometry and the microscopic investigation of the solid phase provides information on the crystallization kinetics [1]. Experimental results under different crystallization conditions (supersaturation, temperature and enantiomeric excess) in a batch crystallizer will be presented. Furthermore, we will show preliminary experiments in which the influence of the counter-enantiomer on the crystal growth of the desired enantiomer is investigated. Common goals for the control of crystallization processes are to obtain a product with prescribed distribution. In the case of preferential crystallization, an additional requirement is the purity of the final product. When preferential crystallization is performed in a cyclic operation mode, the initial state of each cycle depends on the final state of the previous cycle. The difference in the mass fractions of the two enantiomers will vary from cycle to cycle. The control problem then is to determine the optimal temperature cooling profile and the optimal mass of the added seeds and racemate. The concept of orbital flatness constitutes a highly promising approach to this problem [2].
[1] M.P. Elsner, Chirality 2005
[2] U. Vollmer, International Journal of Control 2003, 76
2005
Conference-Paper
http://edoc.mpg.de/237828
Chemie Ingenieur Technik, v.77, 8, 1102 (2005)
en
oai:edoc.mpg.de:2378382005-11-0719:91
Batch crystallisation control based on population balance models
Vollmer, U.
Raisch, J.
expertsonly
In this contribution, we investigate feedforward control synthesis for a batch cooling crystalliser based on a simple population balance model. Population balance models are multidimensional dynamical systems, where one of the independent variables represents time, the other(s) "property coordinate(s)" as, e.g. particle size. They typically describe the temporal evolution of number density functions and are therefore naturally suited to model particulate processes in chemical engineering. Crystallisation represents an important class of particulate processes. It is well established in the chemical and pharmaceutical industry as a purification and separation technique. The quality of crystalline products is strongly influenced by physical properties, such as crystal size distribution (CSD). In batch cooling crystallisers, the fact that solubility depends on temperature is exploited. In this contribution, the process is described by a standard population balance model consisting of a partial differential equation and an ordinary differential equation. The model describes the temporal evolution of both crystal size distribution and solute concentration. Temperature in the crystalliser is considered to be the control input, and the control objective is to achieve a desired crystal size distribution at the end of the batch. The suggested procedure uses a time scaling which transforms the partial differential equation into a simple transport equation and makes the associated (finite dimensional) moment model a flat systems. It checks whether the desired final CSD is achievable, i. e. whether it is compatible with the model assumptions and, if the outcome is affirmative, generates the appropriate control signal.
2004
Conference-Paper
http://edoc.mpg.de/237838
urn:ISBN:90-5682-517-8
Proc. MTNS2004 - 16th International Symposium on Mathematical Theory of Networks and Systems (2004)
en
oai:edoc.mpg.de:2378412005-11-0719:91
Hierarchical Hybrid Control of a Multiproduct Batch Plant
Moor, T.
Raisch, J.
expertsonly
2005
Conference-Paper
http://edoc.mpg.de/237841
Proc. 16th IFAC World Congress (2005)
en
oai:edoc.mpg.de:2378432006-05-0319:91
Hierarchical Hybrid Control Synthesis and its Application to a Multiproduct Batch Plant
Raisch, J.
Moor, T.
expertsonly
This contribution demonstrates the potential of a hierarchical hybrid control framework developed in a previous paper (Moor et al., 2003) by applying it to a specific process control example. The example is the control of a multiproduct batch plant with both continuous and discrete inputs. The control problem is to produce a desired amount of each product with minimal operating costs while observing safety and quality requirements.
2005
Conference-Paper
http://edoc.mpg.de/237843
urn:ISBN:3-540-27938-5
Meurer, T.; Graichen, K.; Gilles, E. D.: Control and Observer Design for Nonlinear Finite and Infinite Dimensional Systems, 199-216 (2005)
en
oai:edoc.mpg.de:2378442012-03-0519:91
On the convexity of reachable sets of nonlinear dynamic systems : an important step in generating discrete abstractions of continuous systems
Geist, S.
Reißig, G.
Raisch, J.
expertsonly
2005
Conference-Paper
http://edoc.mpg.de/237844
urn:ISBN:83-60140-90-1
urn:ISBN:83-60140-85-5
MMAR 2005 - 11th IEEE International Conference on Methods and Models in Automation and Robotics, 101-106 (2005)
en
oai:edoc.mpg.de:2378742005-11-0919:91
Optimal Control of Batch and Semibatch Reactors
Gromov, D.
Mishra, B.
Raisch, J.
expertsonly
The Pontryagin Maximum Principle is a recognised tool for optimal control problems. The main drawback of this approach is that it requires solving a two-point boundary value problem (TBVP) that may be rather involved. This contribution discusses how qualitative methods can considerably reduce the complexity of the resulting TBVP. The obtained results are used for optimal control of batch and semi-batch reactors.
2005
Conference-Paper
http://edoc.mpg.de/237874
Proceedings of the 11th IEEE International Conference on Methods and Models in Automation and Robotics, 107-112 (2005)
en
oai:edoc.mpg.de:2381172012-03-0519:91
Reduction of Chemical Reaction Networks Using Quasi-Integrals
Straube, R.
Flockerzi, D.
Mueller, S. C.
Hauser, M. J. B.
expertsonly
We present a numerical method to identify possible candidates of quasi stationary manifolds in complex reaction networks governed by systems of ordinary differential equations. Inspired by singular perturbation theory we examine the ratios of certain components of the reaction rate vector. Those ratios that rapidly approach an almost constant value define a slow manifold for the original flow in terms of quasi integrals, i.e. functions that are almost constant along the trajectories. The dimensionality of the original system is thus effectively reduced without relying on a priori knowledge of the different time scales in the system. We also demonstrate the relation of our approach to singular perturbation theory which, in its simplest form, is just the wellknown quasi-steady-state approximation. In two case studies, we apply our method to oscillatory chemical systems: the 6-dimensional hemin -- hydrogen peroxide -- sulfite pH -- oscillator and a 10-dimensional mechanistic model for the peroxidase - oxidase (PO) reaction system. We conjecture that the presented method is especially suited for a straight forward reduction of higher dimensional dynamical systems where analytical methods fail to identify the different time scales associated with the slow invariant manifolds present in the system.
2005
Article
http://edoc.mpg.de/238117
Journal of Physical Chemistry A, v.109, 441-450 (2005)
en
oai:edoc.mpg.de:2381182007-09-2019:91
Mathematical Model of Influenza A Virus Production in Large-Scale Microcarrier Culture
Möhler, L.
Flockerzi, D.
Sann, H.
Reichl, U.
expertsonly
mathematical model; virus dynamics; animal cell culture; influenza virus; vaccine production
A mathematical model that describes the replication of influenza A virus in animal cells in large-scale microcarrier culture is presented. The virus is produced in a two-step process, which begins with the growth of adherent Madin-Darby canine kidney (MDCK) cells. After several washing steps serum-free virus maintenance medium is added, and the cells are infected with equine influenza virus (A/Equi 2 (H3N8), Newmarket 1/93). A time-delayed model is considered that has three state variables: the number of uninfected cells, infected cells, and free virus particles. It is assumed that uninfected cells adsorb the virus culture.at the time of infection. The infection rate is proportional to the number of uninfected cells and free virions.Depending on multiplicity of infection (MOI), not necessarily all cells are infected by this first step leading to the production of free virions. Newly produced viruses can infect the remaining uninfected cells in a chain reaction. To follow the time course of virus replication, infected cells were stained with fluorescent antibodies. Quantitation of influenza viruses by a hemagglutination assay (HA) enabled the estimation of the total number of new virions produced, which is relevant for the production of inactivated influenza vaccines. ...
2005
Article
http://edoc.mpg.de/238118
Biotechnology and Bioengineering, v.90, 46-58 (2005)
en
oai:edoc.mpg.de:2381202008-01-1619:91
The Origin of Bursting pH Oscillations in an Enzyme Model Reaction System
Straube, R.
Flockerzi, D.
Mueller, S. C.
Hauser, M. J. B.
expertsonly
bursting oscillations; mixed-mode oscillations; slow-fast analysis; quasi-integrals; two-paramater continuation
The transition from simple periodic to bursting behavior in a 3-dimensional model system of the hemin -- hydrogen peroxide -- sulfite pH oscillator is investigated. A two-parameter continuation in the flow rate and the hemin decay rate is performed to identify the region of complex dynamics. The bursting oscillations emerge subsequent to a cascade of period doubling bifurcations and the formation of a chaotic attractor in parameter space where they are found to be organized in periodic-chaotic progressions. This suggests that the bursting oscillations are not associated with phase-locked states on a 2-torus. The bursting behavior is classified by a bifurcation analysis using the intrinsic slow-fast structure of the dynamics. In particular, we find a slowly varying quasi-species (i.e. a linear combination of two species) which acts as an `internal' or quasi-static bifurcation parameter for the remaining 2-dimensional subsystem. A systematic two-parameter continuation in the internal and one of the external bifurcation parameters reveals a transition in the bursting mechanism from subHopf/fold-cycle to fold/subHopf type. In addition, the slow-fast analysis provides an explanation for the origin of quasi-periodic behavior in the hemin system, even though the underlying mechanism might be of more general importance.
2005
Article
http://edoc.mpg.de/238120
Physical Review E, v.72, 066205 (12 pages) (2005)
en
oai:edoc.mpg.de:2381262012-11-1219:91
Equilibrium and Rate-Based Approaches to Liquid-Liquid Phase Splitting Calculations
Steyer, F.
Flockerzi, D.
Sundmacher, K.
expertsonly
Phase splitting; liquid-liquid equilibrium; nonequilibrium thermodynamics; homotopy continuation
Several approaches are presented that can be used to compute liquid-liquid phase splitting. These approaches are based on homotopy continuation methods or on nonequilibrium thermodynamics. It is shown that they exhibit quite different computational demands with respect to computation times and a suggestion is made as to which algorithm to use. For chemical systems with only a few components the computation times encountered on a normal PC in most cases are only a few milliseconds, in some cases even less -- suggesting the general use of such algorithms to assure that phase splitting is always detected when present.
2005
Article
http://edoc.mpg.de/238126
Computers & Chemical Engineering, v.30, 277-284 (2005)
en
oai:edoc.mpg.de:2399172009-11-3019:91
Using Chemical Reaction Network Theory to discard a Kinetic Mechanism Hypothesis
Conradi, C.
Saez-Rodriguez, J.
Gilles, E. D.
Raisch, J.
expertsonly
Chemical Reaction Network Theory; Model discrimination; MAPK-cascade
Feinberg’s Chemical Reaction Network Theory (CRNT) connects the structure of a biochemical reaction network to qualitative properties of the corresponding system of ordinary differential equations. In particular, no information about parameter values is needed. As such, it seems to be well suited for application in systems biology, where parameter uncertainty is predominant. However, its application in this area is rare, at best. To demonstrate potential benefits from the application, different reaction networks representing a single layer of the well studied Mitogen-activated protein kinase (MAPK) cascade are analyzed. Recent results from Markevich et al. (2004) show that, unexpectedly, multilayered protein kinase cascades can exhibit multistationarity even on a single cascade level. Using CRNT, we show that their assumption of a distributive mechanism for double phosphorylation and dephosphorylation is crucial for multistationarity on the single cascade level.
2005
Article
http://edoc.mpg.de/239917
IEE Proceedings Systems Biology, v.152, 243-248 (2005)
en
oai:edoc.mpg.de:2448252013-02-0819:91
Optimal control of discrete-time hybrid automata under safety and liveness constraints
Seatzu, C.
Gromov, D.
Raisch, J.
Corona, D.
Giua, A.
expertsonly
In this contribution we address an optimal control problem for a class of discrete-time hybrid automata under safety and liveness constraints. The solution is based on a hierarchical decomposition of the problem, where the low-level controller enforces safety and liveness constraints while the high-level controller exploits the remaining degrees of freedom for performance optimisation. Lower-level control is based on a discrete abstraction of the continuous dynamics. The action of low-level control can be interpreted as restricting invariants in the hybrid automaton representing the plant model. A state feedback solution for the high-level control is provided, based on the off-line construction of an appropriate partition of the state space.
Copyright © 2005 Elsevier Ltd. All rights reserved. [accessed February 8th 2013]
2006
Article
http://edoc.mpg.de/244825
Nonlinear Analysis :Theory, Methods & Applications, v.65, 1188-1210 (2006)
en
oai:edoc.mpg.de:2449062009-11-2519:91
Gait phases detection and step length estimation of gait by means of inertial sensors
Negaard, N.-O.
Kauert, R.
Andres, S.
Schauer, T.
Raisch, J.
expertsonly
The application of inertial sensors in the FES-aided gait rehabilitation is investigated. Instead of using foot switches, inertial sensors can be applied to detect gait phases. It is shown that it is possible to reliably detect four distinct gait phases by use of inertial sensors. Additionally, foot orientation is estimated by integration of angular velocity. To avoid drift in the orientation estimate a Kalman filter was implemented taking the measured acceleration in the swing phase into account. The foot position during the swing phase of gait can be estimated on the basis of measured acceleration and already calculated orientation. The accuracy of this calculation is increased by introducing start and stop constraints on the velocity. To validate the algorithms described in this paper, two gait trials with stroke patients were performed on a treadmill. In these trials an ultrasonic measurement unit was used as a reference system. The results show that the step length can be estimated quite accurately. The standard deviation of the errors of the step length estimation compared to the reference measurement were lying between 3 % and 5 %. The gait phases detection and foot position estimate are expected to be usable in FES-aided gait rehabilitation for stroke patients. The reconstructed position and orientation of the foot can be employed as feedback signals for automatic tuning of muscle stimulation in such systems.
International Federation for Medical and Biological Engineering
2005
Conference-Paper
http://edoc.mpg.de/244906
3rd European Medical & Biological Engineering Conference (EMBEC 2005), International Federation for Medical and Biological Engineering (2005)
en
oai:edoc.mpg.de:2449082005-11-0919:91
Control of FES-aided gait by means of inertial sensors
Negard, N.-O.
Schauer, T.
Raisch, J.
expertsonly
Inertial sensors might be used as replacement of traditional foot switches in the control of stimulation in a Functional Electrical Stimulation (FES) aided gait rehabilitation. The inertial sensors can be used to detect gait phases, which in order can be used to synchronise the stimulation to defined gait phases for the different channels. Additionally, inertial sensors can be applied to estimate orientation and 3 dimensional movement of the foot. Based on orientation and estimated 3D curve, several movement parameters can be defined. The most important are foot clearance, which is defined as maximum distance between foot and ground, and sagittal angle of the foot in relation to ground at the time instant close before ground contact. A model of the legs during swing phase was developed on basis of first rinciples and combined with a complex physiological model from literature. The aim of this work was to evaluate the use of movement parameters obtained by inertial sensors in a feedback system. A discrete time, gait-cycle to gait-cycle, PI-type controller was designed to control the foot clearance by use of the hamstring stimulation. The simulation trial showed that the controller was able
to react on a reference change in five cycles.
2005
Conference-Paper
http://edoc.mpg.de/244908
urn:ISBN:3-910102-79-4
Schauer, W.: 4th International Symposium on Automatic Control (AUTSYM 2005), o.S. (2005)
en
oai:edoc.mpg.de:2449112012-11-1719:91
Belastungsregelung bei der Elektrostimulationsergometrie
Schauer, T.
Salbert, R. C.
Negaard, N.-O.
Hunt, K. J.
Raisch, J.
expertsonly
Biomedical engineering; self-tuning control; extended Kalman filter
Für querschnittgelähmte Menschen stellt das Herz-Kreislauf-Training an Fahrradergometern mit Funktioneller Elektrischer Stimulation (FES) der gelähmten unteren Extremitäten eine sich zunehmend etablierende Rehabilitationsmaßnahme dar. Diese Arbeit stellt ein neues Ergometersystem mit Elektrostimulation vor, bei dem die Beine zusätzlich durch einen Hilfsmotor unterstützt werden. Durch den Motor kann ein isokinetisches Training realisiert werden. Der Patient arbeitet in diesem Fall mit den elektrisch stimulierten Muskeln gegen den Motor, der die Trittgeschwindigkeit konstant hält. Für die gezielte Anpassung der Belastung durch die Elektrostimulation wurde ein sich selbst einstellender Momentenregler entwickelt. Der Reglerentwurf erfolgt anhand der online identifizierten linearen Übertragungsfunktion zwischen Stimulationsintensität und gemitteltem muskulären Antriebsmoment. Erste experimentelle Ergebnisse mit einem neurologisch intakten Probanden werden vorgestellt.
2005
Article
http://edoc.mpg.de/244911
At-Automatisierungstechnik, v.53, 607-614 (2005)
de
oai:edoc.mpg.de:2449132005-11-0919:91
Blood glucose control in critical ill patients
Schauer, T.
Ferchland, E.
Raisch, J.
expertsonly
Nonlinear hybrid control of blood glucose in critical ill patients has been investigated. Standard control algorithms for blood glucose only adjust the insulin infusion rate for lowering the blood sugar level. Hypoglycaemic situations are critical is this case since no automatic control action can be lunched. To avoid hypoglycaemia, controller performance is usually chosen low what results in large settling times. In the proposed hybrid control scheme, glucose and insulin infusions are administrated. This allows to track a specied blood glucose pro le exactly. The employed controller is model-based and tested in computer simulations.
2005
Conference-Paper
http://edoc.mpg.de/244913
urn:ISBN:3-910102-79-4
Schauer, W.: 4th International Symposium on Automatic Control (AUTSYM 2005), o.S. (2005)
en
oai:edoc.mpg.de:2449152005-11-2119:91
Funktionelles Handöffnen und -schließen mittels EMG-gesteuerter elektrischer Stimulation
Salbert, R. C.
Schauer, T.
Schmidt, S.
Raisch, J.
expertsonly
Für Schlaganfallpatienten stellt die Wiederherstellung der Greifbewegung mit funktioneller elektrischer Stimulation (FES) der gelähmten Hand eine sich zunehmend etablierende Rehabilitationsmaßnahme dar. Die derzeit existierenden technisch-therapeutischen Lösungen sind jedoch nicht zufriedenstellend. Diese Arbeit stellt einen Elektromyogramm(EMG)-gesteuerten-Zustandsautomaten vor, der die Elektrostimulation in Anhängigkeit von der Patientenintention steuert um das Greifen zu realisieren. Dabei werden die Grundfunktionen Öffnen und Schließen der Hand so kombiniert, dass ein Greifen ermöglicht wird.
2005
Conference-Paper
http://edoc.mpg.de/244915
urn:ISBN:3-910102-79-4
Schauer, W.: 4th International Symposium on Automatic Control (AUTSYM 2005), o.S. (2005)
de
oai:edoc.mpg.de:2449982013-02-0819:91
A hierarchy of discrete abstractions for a hybrid plant (extended version)
Raisch, J.
expertsonly
1998
Article
http://edoc.mpg.de/244998
Journal Europeen des Systemes Automatises = European Journal of Automation, v.32, 1073-1095 (1998)
en
oai:edoc.mpg.de:2449992009-11-1219:91
Discrete approximation and supervisory control of continuous systems
Raisch, J.
O'Young, S. D.
expertsonly
This contribution addresses the following hybrid control problem: a continuous plant (its state evolving in Euclidean space) is to be controlled via symbolic outputfeedback – both measurement and control signal “live” on finite sets of symbols. We adopt the following approach: the hybrid problem is first translated into a purely discrete problem by approximating the continuous plant model by a (nondeterministic) finite state machine. By taking into account past measurement and control symbols, approximation accuracy can be improved and adjusted to the specification requirements. Supervisory control theory for DES (Discrete Event Systems) is then applied to find the optimal controller which enforces the specifications. As the behaviour of the approximating automaton is guaranteed to contain the behaviour of the underlying continuous plant model, the controller also forces the latter to obey the specifications.
1998
Article
http://edoc.mpg.de/244999
IEEE Transactions on Automatic Control, v.43, 569-573 (1998)
en
oai:edoc.mpg.de:2451642013-02-0819:91
Control of Batch Crystallization : a System Inversion Approach
Vollmer, U.
Raisch, J.
expertsonly
batch crystallization; population balance; process control; orbital flatness
In this article a new approach to the control of batch crystallization is presented. In batch cooling crystallization the crucial control problem is to design a temperature trajectory which produces a desired crystal size distribution at the end of the batch. Traditionally, this problem is addressed in an optimization framework. Here, a completely different solution is presented. It is shown that the standard population balance model can be inverted analytically. The system inversion is performed making use of a state dependent time scaling of the population balance model. Consequently, for any desired crystal size distribution the corresponding temperature trajectory (i.e. the feed-forward control) can be determined in a straight forward way. Furthermore, exploiting the properties of the time-scaled moment model, a nonlinear feedback control is designed for the batch crystallizer, which ensures tracking of the previously designed trajectories in the presence of uncertainty.
Copyright © 2006 Elsevier B.V. All rights reserved. [accessed February 8th 2013]
2006
Article
http://edoc.mpg.de/245164
Chemical Engineering & Processing, v.45, 874-885 (2006)
en
oai:edoc.mpg.de:2451682013-02-0819:91
Learning by Doing : Systematic Abstraction Refinement for Hybrid Control Synthesis
Moor, T.
Davoren, J. M.
Raisch, J.
expertsonly
2006
Article
http://edoc.mpg.de/245168
IEE Proceedings - Control Theory and Applications, v.153, 591-599 (2006)
en
oai:edoc.mpg.de:2451692012-03-2719:91
Throughput-Optimal Sequences for Cyclically Operated Plants
Mayer, E.
Haus, U.-U.
Raisch, J.
Weismantel, R.
expertsonly
cyclic scheduling; high throughput screening; integer programming
2008
Article
http://edoc.mpg.de/245169
Discrete Event Dynamic Systems - Theory and Applications, v.18, 355-383 (2008)
en
oai:edoc.mpg.de:2451722005-11-1019:91
On suboptimal control design for hybrid automata using predictive control techniques
Spathopoulos, M.
Pang, Y.
Raisch, J.
expertsonly
2005
Conference-Paper
http://edoc.mpg.de/245172
Proceedings of the 16th IFAC World Congress (2005)
en
oai:edoc.mpg.de:2459902005-11-2119:91
Control problems in batch crystallisation of enantiomers
Raisch, J.
Vollmer, U.
Angelov, I.
expertsonly
In the first part of this contribution, we investigate control synthesis for a batch cooling crystalliser based on a simple population balance model. Population balance models are multidimensional dynamical systems, where one of the independent variables represents time, the other(s) "property coordinate(s)" as, e.g., particle size. They typically describe the temporal evolution of number density functions and are therefore naturally suited to model particulate processes in chemical engineering as, e.g., crystallisation. We demonstrate that the corresponding moment model is orbitally flat, i.e., it can be transformed into a flat system by an appropriate time transformation. Roughly speaking, flat systems possess an invertibility property, and this can be conveniently used to generate an input signal that will result in a desired crystal size distribution at the end of the batch. In the second part of this contribution, we indicate how the described flatness based approach can be extended to address the more demanding problem of enantiomer separation by crystallisation.
Oprogramowanie Naukowo-Techniczne
2005
Conference-Paper
http://edoc.mpg.de/245990
urn:ISBN:83-916420-3-8
Tadeusiewicz, R.; Liegeza, A.; Szymkat, M.: CMS'05 - Computer Methods and Systems : Plenary lectures and special session papers; Volume I, Oprogramowanie Naukowo-Techniczne, 189-194 (2005)
en
oai:edoc.mpg.de:2934792013-02-0819:91
Optimal Operation of Enantioseparation by Batch-Wise Preferential Crystallization
Angelov, I.
Raisch, J.
Elsner, M. P.
Seidel-Morgenstern, A.
expertsonly
Preferential crystallization; Batch-wise non-isothermal crystallization; Dynamic optimization; B-spline interpolation
This article describes batch-wise preferential crystallization separation of mixtures of L- and D- threonine. Use of online polarimetry combined with refractometry and microscopic investigation of the solid phase provides information on the crystallization kinetics. Results obtained for different crystallization conditions (supersaturation, temperature and enantiomeric excess) in a batch crystallizer are presented. Based on these results, a nonlinear dynamic model has been developed. The control problem is to determine an optimal temperature profile which will result in a maximum amount of product with required quality. In this dynamic optimization problem B-splines have been used for interpolation of the temperature profile.
Copyright © 2007 Elsevier Ltd All rights reserved.
[accessed June 6, 2008]
2008
Article
http://edoc.mpg.de/293479
Chemical Engineering Science, v.63, 1282-1292 (2008)
en
oai:edoc.mpg.de:2934862012-11-1219:91
Adaptive high-resolution schemes for multidimensional population balances in crystallization processes
Qamar, S.
Ashfaq, A.
Warnecke, G.
Angelov, I.
Elsner, M. P.
Seidel-Morgenstern, A.
expertsonly
This article focuses on the application of adaptive high-resolution finite volume schemes for solving multidimensional population balance models (PBM) in crystallization processes. For the mesh redistribution, we use the moving mesh technique of Tang and Tang [Tang, H.-Z. & Tang, T. (2003). Adaptive mesh methods for one- and two-dimensional hyperbolic conservation laws. SIAM Journal of Numerical Analysis, 41, 487 515] which they have developed for hyperbolic conservation laws in conjuction with finite volume schemes. In this technique, an iterative procedure is used to redistribute the mesh by moving the spatial grid points. The corresponding numerical solution at the new grid points is obtained by solving a linear advection equation. The method avoids the usual unsatisfactory, interpolation procedure for updating the solution. The finite volume schemes were originally derived for compressible fluid dynamics. The schemes have already shown their accuracy and efficiency in resolving sharp peaks and shock discontinuities. The accuracy of these schemes has been improved further by using the adaptive meshing techniques. The application of these high-resolution schemes for multidimensional crystallization processes demonstrates their generality, efficiency, and accuracy. The numerical test cases presented in this article show the clear advantage of finite volume schemes and show further improvements when combined with amoving mesh technique.
2007
Article
http://edoc.mpg.de/293486
Computers and Chemical Engineering, v.31, 1296-1311 (2007)
en
oai:edoc.mpg.de:2935402007-03-2019:91
A gradient-based approach to a class of hybrid optimal control problems
Azhmyakov, V.
Raisch, J.
expertsonly
We investigate optimal control problems for a class of non-stationary hybrid systems with autonomous location transitions.Using the Lagrange approach and the technique of the reduced gradient, we derive necessary optimality conditions for the considered class of problems. These optimality conditions are closely related to a variant of the Hybrid Maximum Principle and can be used for constructive optimization algorithms.
2006
Conference-Paper
http://edoc.mpg.de/293540
Proceedings of the 2nd IFAC Conference on Analysis and Design of Hybrid Systems (ADHS), 89-94 (2006)
en
oai:edoc.mpg.de:2978902007-02-0619:91
Detecting and enforcing monotonicity for hybrid control systems synthesis
Gromov, D.
Raisch, J.
expertsonly
Abstraction based approaches to control of hybrid systems require efficient means of computing outer approximations of reachable continuous state sets. This contribution discusses how the concept of monotonicity can be used for this purpose. It provides an efficient algorithm to check whether a given continuous system is monotone with respect to a (a-priori unknown) partial order and, if not, investigates how to use continuous feedback to enforce monotonicity. In the latter case, the resulting continuous feedback represents a (lower) control level within a hierarchical hybrid control system.
2006
Conference-Paper
http://edoc.mpg.de/297890
Proceedings of the 2nd IFAC Conference on Analysis and Design of Hybrid Systems (ADHS), 395-402 (2006)
en
oai:edoc.mpg.de:2980222007-02-0619:91
Timed Discrete Event Control of a Parallel Production Line with Continuous Output
Gromov, D.
Geist, S.
Raisch, J.
expertsonly
In this paper we present an approach to formulate and solve certain scheduling tasks using timed discrete event control methods. To demonstrate our approach, we consider a special class of systems: a cyclically operated chemical plant with parallel reactors using common resources and a continuous output. This problem was motivated by a benchmark proposed within the EU Network of Excellence HYCON. For this class of systems, we show how to pose the control problem within a discrete event framework by modelling system components as multirate timed automata. Safety and nonblocking are investigated. These properties have to be achieved in the presence of a class of bounded errors/disturbances.
2006
Conference-Paper
http://edoc.mpg.de/298022
Proceedings of the 2nd IFAC Conference on Analysis and Design of Hybrid Systems (ADHS), 205-211 (2006)
en
oai:edoc.mpg.de:2981782007-02-0919:91
Chemical Reaction Network Theory : a tool for systems biology
Conradi, C.
Saez-Rodriguez, J.
Gilles, E. D.
Raisch, J.
expertsonly
Feinberg's Chemical Reaction Network Theory (CRNT) connects the structure of a biochemical reaction network to qualitative properties of the corresponding system of ordinary differential equations. In particular, no information about parameter values is needed. As such, it seems to be well suited for application in systems biology, where parameter uncertainty is predominant. To demonstrate potential benefits from the application, different reaction networks representing a single layer of the well-studied Mitogen-activated protein kinase (MAPK) cascade are analyzed. Recent theoretical results show that multilayered protein kinase cascades can exhibit multistationarity even on a single cascade level. Using CRNT, we examine the influence of conserved moieties on the existence of multiple steady states.
2006
Conference-Paper
http://edoc.mpg.de/298178
Proceedings of the 5th MATHMOD : 5th Vienna Symposium on Mathematical Modelling (2006)
en
oai:edoc.mpg.de:2981852012-04-1119:91
Regelung und Diagnose der Elektrotherapie mittels Elektromyographie am stimulierten Muskel
Salbert, R. C.
Schauer, T.
Hesse, S.
Bardeleben, A.
Raisch, J.
expertsonly
2006
Conference-Paper
http://edoc.mpg.de/298185
Tagungsband des Workshops Automatisierungstechnische Methoden und Systeme in der Medizin : Automed 2006 (2006)
de
oai:edoc.mpg.de:2981872012-03-0219:91
Hybride modell-prädiktive Regelung des Blutzuckerspiegels auf Intensitivstationen
Schauer, T.
Goerner, M.
Raisch, J.
expertsonly
2006
Conference-Paper
http://edoc.mpg.de/298187
Tagungsband des Workshops Automatisierungstechnische Methoden und Systeme in der Medizin : Automed 2006 (2006)
de
ResultSet_820C5lViaox_range_100-199