Performance Judgment of Automotive Wire Harness Based on Convolutional Neural Network

This paper describes a performance judgment method based on a convolutional neural network (CNN) for an automotive wire harness. The proposed method uses the CNN to represent the correlation between the cross-sectional shape of the wire harness and its electric performance. We consider combinations of a few patterns of the inputs and outputs of the CNN and discuss about the usefulness of each input-output pattern for the proper judgment.