Neuro-fuzzy systems for engineering applications
K. Goser, Karlheinz Schuhmacher, Matthias Härtung, Kai Heesche, Benjamin Hesse, Andreas Kanstein · 2002
This paper reports three advanced developments in neuro-fuzzy systems for engineering applications. The essential part of neuro-fuzzy synergism's comes from a common framework called adaptive networks, which unifies both neural networks and fuzzy systems. A neural network with the interpretation done by a fuzzy system is suitable for quality monitoring of mechanical equipment or power electronics. Neural networks like self-organizing maps or radial basis function nets can feed the data for fuzzy systems which can control battery chargers or drilling machines. An integrated neuro-fuzzy topology is described which represents both a complete feed forward network and a fuzzy controller. This data driven system is very suitable for automotive control. Neuro-fuzzy systems are an important area for technology transfer. For this transfer the concept at the University of Dortmund is presented as a successful model.