Subproject C3 is working towards application-oriented development of methods, procedures and structures for control systems of self-optimizing mechatronic systems. In this effort concepts of mathematics, control theory and computer science are joined and made usable for self-optimization. The self-optimization process is of special interest in a modular and hierarchical system with the OCM (operator-controller-module) as a central structuring element. Model-based and behaviour-based methods are used along with planning-based methods.
In subproject C3 different types of methods for realization of self-optimizing mechatronic systems are explored. One important focus is on the analysis of active structures in hierarchically coupled systems. The resulting concept is based on a hierarchical model of the system that is used to implement a knowledge base that is distributed over the hierarchy. Information exchange between the single substructures is accomplished using reduced models. These hierarchical models are the basis for the development of hierarchical optimization techniques.
|
 |
Approach of hierarchical optimi- zation in the cognitive oper- ator of the OCM. The optimi- zation of the higher level module chooses points from the pareto set of the lower level mo- dule. The changed be- haviour of the lower level be- haviour is communicated back as reduced behaviour models. |
The process of self-optimization works inside the single substructures and is characterized by a three step process: Analyzing the current situation, determining the system objectives and adaptation of the system behaviour. The following text will assign the methods analyzed in this subproject to these three steps.
As part of the analysis of the current situation, information about the current situation of the system and its environment is gathered. The realization of this step uses identification methods using classic methods as well as iterative learning. The current situation of the system and its environment can be captured using observers and classic identification methods. Observers are used to estimate the current system state as well as disturbances of the environment. Identification methods are used to determine changes of system parameters. Environmental states can also be acquired by communication with other systems. Future repetitive disturbances can be captured using iterative learning methods and globally distributed this way. Information acquired this way can be stored locally in a knowledge base or in a global knowledge base for shared use with other systems. The following steps of the self-optimization process can access them too.
The step of determination of objectives uses information of the knowledge base to create new objectives through selection or adaptation. For realization of such an objective, two approaches are considered: monitoring of dependent variables and hybrid planning. Self-optimizing systems always have to deal with differences between nominal and actual values of dependent variables because of unconsidered or unknown influence factors. The monitoring of dependent variables “reacts” upon this difference and adapts the system of objectives analogous to a controller. Based on the knowledge of known future system and environment state the hybrid planner can determine the upcoming state of the system of objectives. A plan is created that allows the system to act upfront an event.
In the step of adaptation of the system behaviour the optimal system settings are determined using online-optimization represented by pareto sets that are stored in the knowledge base. Parameters from the pareto set are read using the current system of objectives and are activated in the system. The system settings are deployed in the controller system by changing parameters of structure. Concepts of systematic switch-over for controller structures are developed to support this. Gradients are calculated using algorithmic differentiation to support and speed up the online-optimization.
Safety aspects are of paramount importance in self-optimizing systems. Ongoing work in subproject C3 will address system behaviour in presence of aperiodic disturbances or those not exactly known or fluctuating as well as loss of single units. The approach will consider uncertainty and loss of actuators in the whole self-optimization process by choosing system configurations that keep a certain notion of optimality under these conditions.
Publications (since 6/2005)
Reviewed Publications
Adelt, P.; Schmidt, A.; Esau, N.; Kleinjohann, L.; Kleinjohann, B.; Rose, M.: Approximation of environment models for an air gap adjustment system in a hybrid planning context. In: IEEE International Symposium on Intelligent Control (ISIC), September 17-20, 2008, Vilnius, IEEE Control Systems Society, San Antonio, Texas, USA, 2008
Burmester, S.; Giese, H.; Henkler, S.; Hirsch, M.; Tichy, M.; Gambuzza, A.; Münch, E.; Vöcking, H.: Tool Support for Developing Advanced Mechatronic Systems: Integrating the Fujaba Real-Time Tool Suite with CAMeL-View. In: Proceedings of the 29th International Conference on Software Enigneering (ICSE), May 20-26, 2007, Minneapolis, IEEE Computer Society Press, 2007, pp. 801-804
Burmester, S.; Giese, H.; Münch, E.; Oberschelp, O.; Klein, F.; Scheideler, P.: Tool Support for the Design of Self-Optimizing Mechatronic Multi-Agent Systems. International Journal on Software Tools for Technology Transfer (STTT), 8 (4), Springer Verlag, Berlin / Heidelberg, 2008
Çinkaya, H.; Münch, E.; Nölle, R.; Trächtler, A.: Self-Calibration of the 6-dof Parallel Robot TriPlanar by Identification of the Geometry Parameters. In: IEEE/ASME International Conference on Advanced Intelligent Mechatronics (AIM), September 4-7, 2007, Zürich, 2007
Çinkaya, H.; Münch, E.; Nölle, R.; Trächtler, A.; Toepper, S.: Selbstkalibrierung des 6 DOF Parallelroboters TRIPLANAR durch Identifikation der Geometrieparameter mit einem einfachen Messverfahren. In: Mechatronik 2007 - Innovative Produktentwicklung, 23.-24. Mai 2007, Wiesloch, VDI-Berichte, Band 1971, 2007
Esau, N.; Adelt, P.; Schmidt, A.: Interval Optimization in a Hybrid Planner Context for an Air Gap Adaptation System. In: Proceedings of the 3rd Asia International Symposium on Mechatronics (AISM), August 27-31, 2008, Sapporo, 2008
Gambuzza, A.; Koert, D.: A Concept for a Platform-Independent Modelling of Mechatronic Systems: Improving the Re-usability of Models in the Mechatronic Design Process. In: Postworkshop Proceedings of the 3rd Workshop on Object-oriented Modeling of Embedded Real-Time Systems (OMER), October 12-14, 2005, Paderborn, 2005
Gambuzza, A.; Koert, D.; Münch, E.; Vöcking, H.: Automatische Codegenerierung für verteilte Informationsverarbeitung in mechatronischen Systemen. Tagungsband des 4. Paderborner Workshop - Entwurf mechatronischer Systeme, 30.-31. März 2006, Paderborn, HNI-Verlagsschriftenreihe, Band 189, 2006
Gausemeier, J.; Radkowski, R.; Kleinjohann, B.; Richert, W.; Zabel, H.and Adelt, P.: Augmented Reality for the Testing of Autonomous Systems by the Example of Soccer Robots - Concepts and Experimental Results. In: Workshop Virtual Prototyping Applications, International Scientific Colloquium (IWK), September 8-12, 2008, Ilmenau, 2008
Koch, M.; Beckebans, R.; Schrage, J.: Movement reconstruction in industrial and medical domains based on topological Petri nets. In: Proceedings of 6th International Industrial Simulation Conference (ISC 2008), EUROSIS-ETI, Lyon, France, 2008
Koch, M.; Beckebans, R.; Schrage, J.; Richert, W.: Real-time measurement, visualization and analysis of movements by fiber optical sensory applied to robotics. In: Proceedings of 47th International IEEE Conference on Instrumentation, Control and Information Technology (SICE 2008), SICE, Tokyo, Japan, 2008
Koch, M.; Schäfers, L.; Kleinjohann, B.: An efficient dataflow-oriented fuzzy library. In: Proceedings of 47th International IEEE Conference on Instrumentation, Control and Information Technology (SICE 2008), SICE, Tokyo, Japan, 2008
Koch, M.; Schrage, J.; Richert, W.: Optic-Tactile robotics and medical applications. In: Proceedings of International IEEE/ASME Conference on Advanced Intelligent Mechatronics (AIM 2008), IEEE/ASME, Xi'an, China, 2008
Lückel, J.; Münch, E.; Vöcking, H.; Hestermeyer, T.: Online Optimization of a Preview Controller - Structure and Algorithms. Journal of Theoretical and Applied Mechanics (JTAM), 43, Polish Society of Theoretical and Applied Mechanics, Warschau, 2005, pp. 575-591
Münch, E.; Adelt, P.; Krüger, M.; Kleinjohann, B.; Trächtler, A.: Hybrid Planning and Hierarchical Optimization of Mechatronic Systems. In: 2008 International Conference on Control, Automation and Systems, ICROS, Seoul, Korea, 2008
Münch, E.; Çinkaya, H.; Trächtler, A.: Realisierung eines Werkzeuges zum algorithmischen Differenzieren von Modellen mechatronischer Systeme. 5. Paderborner Workshop Entwurf mechatronischer Sytsteme, HNI-Verlagsschriftenreihe, Band 210, Paderborn, 2007
Münch, E.; Gambuzza, A.; Paiz, C.; Pohl, C.; Porrmann, M.: FPGA-in-the-Loop Simulations with CAMEL-View. Proceedings of the 7th International Heinz Nixdorf Symposium, HNI-Verlagsschriftenreihe, Band 223, Paderborn, 2008
Münch, E.; Trächtler, A.: Observation of Gradients, by Means of Algorithmic Differentiation. In: 32nd Annual Conference on the IEEE Industrial Electronics Society (IECON'06), IEEE IES, Paris, France, 2006, pp. 4743-4748
Münch, E.; Vöcking, H.; Hestermeyer, T.: Self-learning disturbance compensation for active suspension systems. In: Proceedings of the Second International Conference on Informatics in Control, Automation and Robotics (ICINCO 2005), INSTICC Press, Barcelona, Spain, 2005, pp. 32-38
Osmic, S.; Münch, E.; Trächtler, A.; Henkler, S.; Schäfer, W.; Giese, H.; Hirsch, M.: Safe Online-Reconfiguration of Self-Optimizing Mechatronic Systems. Proceedings of the 7th International Heinz Nixdorf Symposium, HNI-Verlagsschriftenreihe, Band 223, Paderborn, 2008
Osmic, S.; Trächtler, A.: Flatness-based Online Controller Reconfiguration. In: 34nd Annual Conference of the IEEE Industrial Electronics Society (IECON'08), IEEE IES, Orlando, Florida, USA, 2008
Schmidt, A.; Adelt, P.; Esau, N.; Kleinjohann, L.; Kleinjohann, B.: Using Environmental Models Approximated by Fuzzy Identification for Hybrid Planning of Mechantronic Systems. In: Proceedings of the 28th Computers and Information in Engineering Conference (CIE), ASME, New York, 2008
Schmidt, A.; Esau, N.; Adelt, P.: Dynamic optimization of track sectioning in mechatronic systems for a hybrid planner. In: Proceedings of the 2008 IEEE International Conference on Mechatronics and Automation (ICMA'08), Takamatsu, Kagawa, Japan, IEEE Computer Society Press, 2008
Schmidt, A.; Esau, N.; Adelt, P.: Integrating dynamic track sectioning into a hybrid planning infrastructure. In: Proceedings of the 2008 SICE International Conference on Instrumentation, Control and Information Technology, SICE, Chofu, Tokyo, Japan, 2008
Trächtler, A.; Münch, E.; Vöcking, H.: Iterative Learning and Self-Optimization Techniques for the Innovative Railcab-System. In: 32nd Annual Conference on the IEEE Industrial Electronics Society (IECON'06), IEEE IES, Paris, France, 2006, pp. 4683-4688
Ph.D.-Theses
Hestermeyer, T.: Strukturierte Entwicklung der Informationsverarbeitung für die aktive Federung eines Schienenfahrzeugs. Dissertation, Fakultät für Maschinenbau, Universität Paderborn, 2006
Oberschelp, O.: Strukturierter Entwurf selbstoptimierender mechatronischer Systeme. Dissertation, Fakultät für Maschinenbau, Universität Paderborn, Vorgelegte Dissertationsschrift, 2008
