Researchers at Purdue University are using a neural network to identify noise and vibration problems on automobiles. A large part of the program involves characterizing the noise and vibration signals from various sub-systems, such as exhaust system, power train, suspension etc, with their associated faults such as floor vibration, cavity boom etc. The work is being funded by ArvinMeritor, a large automotive systems and parts supplier.
Read more on the Purdue University web site.