Using Cased-Based Reasoning Domain Knowledge to Train a Back Propagation NeuralNetwork in order to Classify Gear Faults in an Industrial Robot
Erik Martin Gustaf Olsson · KTH Publication Database DiVA (KTH Royal Institute of Technology) · 2008
The classification performance of a back propagation neural network classifier highly depends on itstraining process. In this paper we use the domain knowledge stored in a Case-based reasoning system inorder to train a back propagation neural network to classify gear faults in an industrial robot. Ourapproach is to compile domain knowledge from a Case-based reasoning system using attributes frompreviously stored cases. These attributes holds vital information usable in the training process. Ourapproach may be usable when a light-weight classifier is wanted due to e.g. lack of computing power orwhen only a part of the knowledge stored in the case base of a large Case-based reasoning system isneeded. Further, no use of the usual sensor signal classification steps such as filtering and featureextraction are needed once the neural network classifier is successfully trained.