Community Research and Development Information Service - CORDIS

ERC

SNLSID Report Summary

Project ID: 320378
Funded under: FP7-IDEAS-ERC
Country: Belgium

Final Report Summary - SNLSID (Data Driven Structured Modelling of Nonlinear Dynamic Systems)

Today’s state-of-the art methods for system and control design are model based. The ever-increasing demand for higher performance and efficiency pushes the systems in a nonlinear operation mode so that nonlinear models are required for their design and control. The model quality and the model building cost are limiting factors for further technological developments.

To close the gap between the designers and the modelers we developed a data driven nonlinear system identification approach to deliver highly structured nonlinear models meeting the designer’s needs. The whole modeling process is considered, from experiment design over model structure selection, to matching the model to the data, and validation of the final results.

From a theoretical point of view, the major contribution is the development of a new unifying approach to the nonlinear data driven modelling problem. The results are centered around the decoupling of multivariate nonlinear functions so that highly structured models are delivered that combine the advantages of sparse representations with an increased intuitive insight in the behavior of the nonlinear system.
From a practical point of view, the new nonlinear modelling paradigm is an enabling technology to further push the performance and efficiency of system and control design.

This new methods allow us to solve today modeling problems that were far beyond the abilities at the start of the project.
Structured nonlinear model building has applications in traditional industrial and emerging new high technological applications coming, amongst others, from the mechanical, electrical, electronic, telecommunication, and automotive field. Also biomechanical and biomedical applications take full advantage of the new modelling strategy. Highly structured nonlinear models provide designers with (intuitive) insight that can guide them towards better solutions for tomorrow’s products.

Reported by

VRIJE UNIVERSITEIT BRUSSEL
Belgium
Follow us on: RSS Facebook Twitter YouTube Managed by the EU Publications Office Top