Neuronale Netze zur Erkennung von Verbrennungsaussetzern im Kraftfahrzeug

Engine misfires can damage a catalytic converter in a short period of time and increase emissions. Thus, the need arises to detect engine misfiring. There are several different methods for detection. This article focuses on the use of neural networks for misfire recognition. The specific requirements that a suitabel neural network has to satisfy for this task will be presented. The software tool MEKS (Feature extraction and classification system) developed by us is described. Current results are presented