Applied Research – Other Topics






Embedded Software

[top]

Visual Programming in the Field of Measurement and Control

In the field of measurement and control, various parameterized algorithms such as digital filters, Fourier transforms, PID controllers, fuzzy classifiers, and so on are typically combined in different ways to solve complex application problems. Thus, a component-based approach is a natural way to support software development in this application domain. Components that encapsulate certain algorithms interact with other components and with the environment of the component system (a technical process, for instance) via well-defined interfaces. Whenever a component produces data according to its algorithm, results are transmitted to the environment or to other components that have to execute their algorithms using those results as inputs. The overall component system may, therefore, be interpreted as a dataflow graph (signal graph).



Signal graph for CRT classification



Examples for results: CRT not defect (left) CRT with crack (right)

Together with Micro-Epsilon Messtechnik and FORWISS Passau we laid the foundations for ICONNECT, a software tool for visual, component-based programming of embedded PC applications. We also realized some applications in the field of measurement and control and integrated neural networks and fuzzy systems into ICONNECT. Together with Micro Epsilon Messtechnik, FORWISS Passau, Jumptec (now: Kontron Embedded Modules), Sedlbauer, and the University of Applied Sciences in Deggendorf (A. Grzemba), we developed an intelligent information system based on ICONNECT and integrated techniques for the combination of data-flow-oriented and control-flow-oriented specification as well as new scheduling algorithms and communication methods.



Further Information

Staff:

Publications:

W. Maydl, B. Sick, W. Grass; Component-Based Measurement and Control Software for Embedded Systems; in: K.-K. Lau (Ed.): Component-Based Software Development: Case Studies; ch. 7, pp. 159-186; Series on Component-Based Software Development, vol. 1; World Scientific Publishing, Singapore; 2004


W. Maydl, B. Sick, W. Grass; Towards a Specification Technique for Component-Based Measurement and Control Software for Embedded Systems; in: M. Fernandez, I. Crnkovic, G. Fohler, C. Griwodz, T. Plagemann, P. Gruenbacher (Eds.): Proceedings of the ”28th EUROMICRO Conference (EUROMICRO 2002)”; pp. 74-80; Dortmund, 2002


A. Sicheneder, A. Bender, E. Fuchs, R. Mandl, M. Mendler, B. Sick; Tool-supported Software Design and Program Execution for Signal Processing Applications Using Modular Software Components; in: T. Margaria, B. Steffen (Eds.): Proceedings of the “International Workshop on Software Tools for Technology Transfer (STTT ’98)”; pp. 61-70; Aalborg, 1998; (BRICS Notes Series NS-98-4, Aarhus/Aalborg)


A. Sicheneder, A. Bender, E. Fuchs, R. Mandl, B. Sick; A Framework for the Graphical Specification and Execution of Complex Signal Processing Applications; in: Proceedings of the “1998 IEEE International Conference on Acoustics, Speech, and Signal Processing (ICASSP ’98)”; vol. 3, pp. 1757-1760; Seattle, 1998


and others...

Patent:

Collaboration:




Various Applications of Computational Intelligence

[top]

Technical Data Mining

Methods from the field of Computational Intelligence and related fields such as Data Mining or Machine Learning can successfully be applied to various technical problems. In the past, we used neural networks, evolutionary algorithms, support vector machines, fuzzy classifiers and controllers, etc. in many application areas to develop novel kinds of “Intelligent Technical Systems”, e.g.,

  • tool condition monitoring (NC lathes),
  • optimization of production processes (e.g., identification of optimal working points of chemical reactors and classification of defects on silicon wafers),
  • wave-front Reconstruction with a Shack-Hartmann Sensor

Many partners were involved in this research, e.g., Weiler, Kistler, Siemens, Brankamp, Agfa-Gevaert, Daimler Benz Aerospace, Rodenstock, Deckel-Maho, Kennametal Hertel, Fraunhofer IIS-A, Fraunhofer IIS-B, Wacker/Siltronic, Micro-Epsilon, FORWISS Passau, and others.

Here are the Lions

“Hic sunt leones!”. In ancient maps from the Roman era, regions about which there was no information were identified by this Latin inscription which can be translated with “Here are the lions”. The reason for the inscription was to warn a traveler about dangers that he may encounter in these regions of the world. We also left our terra cognita from time to time to investigate some business applications of computational intelligence. With Daimler-Chrysler (now: Daimler) we investigated a problem in the field of customer relationship management (CRM) – a ranking of potential customers based on micro-geographic information. We solved this problem by means of evolutionary optimization of radial basis function neural networks. The goal was to select promising addressees for a direct marketing campaign. With NRCM we investigated the use of support vector machines based on dynamic kernel functions to forecast financial time series. The techniques were later integrated into the analysis software Investox.



Investox: Time series forecasting based on support vector machines.



Further Information

Staff:

Publications:

M. Bauer, O. Buchtala, T. Horeis, R. Kern, B. Sick, R. Wagner; Technical Data Mining with Evolutionary Radial Basis Function Classifiers; in: Applied Soft Computing; vol. 9, no. 2, pp. 765-774; 2009


J. Mager, U. Paasche, B. Sick; Forecasting Financial Time Series with Support Vector Machines Based on Dynamic Kernels; in: Proceedings of the ”2008 IEEE Conference on Soft Computing in Industrial Applications (SMCia/08)”; pp. 252-257; Muroran, 2008


D. Fisch, A. Hofmann, V. Hornik, I. Dedinski, B. Sick; A Framework for Large-Scale Simulation of Collaborative Intrusion Detection; in: Proceedings of the ”2008 IEEE Conference on Soft Computing in Industrial Applications (SMCia/08)”; pp. 125-130; Muroran, 2008


O. Buchtala, M. Klimek, B. Sick; Evolutionary Optimization of Radial Basis Function Classifiers for Data Mining Applications; in: IEEE Transactions on Systems, Man, and Cybernetics – Part B: Cybernetics; vol. 35, no. 5, pp. 928-947; 2005


P. Neumann, D. Arndt, B. Sick; An Application of Evolutionary and Neural Data Mining Techniques to Customer Relationship Management; in: M. Kantardzic, J. Zurada (Eds.): Next Generation of Data-Mining Applications; ch. 4, pp. 81-100; Wiley-IEEE Press, New York, Piscataway; 2005


B. Sick; Online and Indirect ToolWear Monitoring in Turning with Artificial Neural Networks: A Review of More Than a Decade of Research; in: Mechanical Systems and Signal Processing; vol. 16, no. 4, pp. 487-546; 2002


B. Sick; Fusion of Hard and Soft Computing Techniques in Indirect, Online Tool Wear Monitoring; in: IEEE Transactions on Systems, Man, and Cybernetics – Part C: Applications and Reviews, Special Issue on Fusion of Soft Computing and Hard Computing in Industrial Applications; vol. 32, no. 2, pp. 80-91; 2002


and others...

Awards:

Collaboration (Selection):