Adaptive Modeling and Pattern Recognition

The Adaptive Modelling and Pattern Recognition group focuses on the development and application of machine learning techniques, innovative algorithms for pattern recognition and modal analysis within funding projects and industrial orders.

Fields of application and research topics are diverse and extensive including mathematical methods for optimization, pattern recognition, prediction, signal processing and sound analysis applied to concrete examples and modifiable by informatics tools.


  • Pattern recognition, classification, optimization, prediction
  • Big data mining, modelling adaptive systems
  • Structural dynamics, vibration analysis, modal analysis,
    operating deflection shapes, transfer path analysis
  • Implementation of innovative barcode technologies

Research Projects


Modal analysis of buildings under varying conditions, e.g. temperature or humidity

With the current operational modal analysis, minor damage to structure is often not recognized at all or discovered too late to be treated properly. Therefore, the project focused on the detection of damages, e.g. caused by changes in temperature or humidity, and the improvement of the modal analysis by means of wavelet transformations.

Funded by: Federal Ministry for Economic Affairs and Energy / INNO-KOM-Ost (MF); funding code: MF120130


Structural analysis of buildings with a programmable shaker and wireless, electromagnetically compatible sensor network 

Development of a comprehensive modular IT system for modal analysis and structural health analysis. Essential part of the project was the optimization and development of algorithms for experimental modal analysis.

Funded by: Federal Ministry for Economic Affairs and Energy / AIF-Projekt GmbH ZIM-KF; funding code: KF2083620KM3


Data and forecast based generation of model parameters for mechanical joints under crash loads

The aim of this research project is firstly to develop a forecasting capability for all mechanical joining technologies to determine joint properties of unknown joints. Experimental and predicted data serve as a base for a finite element model for the crash simulation of mechanical joints. Secondly, the extension of an existing finite element model for self-piercing rivets to other mechanical joining technologies is provided. Another point is the development of an approach for the automated calibration of a finite element model for crash simulation using experimental and prediction based data. The validation of the forecasts of joint properties and the finite element model for crash simulation are performed by using experiments and simulation of specimen and component tests.

Funded by: IGF (AiF e.V., Federal Ministry for Economic Affairs and Energy); funding code: IGF-Nr. 18468BG


Innovative 3D color barcode

A new color barcode that allows individual design in color and shape by integrating the code pattern directly into existing color images and graphics. The free choice of design, shape and colors allow a wide range of coding templates. The unique coding pattern can be recognized easily and distinguishes coded from uncoded images.

Funded by: Federal Ministry for Economic Affairs and Energy / INNO-KOM-Ost (MF); funding code: MF140057


Real-time vibration monitoring system for android devices

Aim of the project was to simplify the mobile analysis and monitoring of vibrations by using an intelligent sensor network. Further, the development of innovative algorithms for data compression in the vibration analysis or monitoring application under real conditions (Big Sensor Data) with an android device.

Funded by: Federal Ministry for Economic Affairs and Energy / INNO-KOM-Ost (MF); funding code: MF150074