Defective judgment for automotive wire harness using convolutional neural network

抄録

This letter proposes a defective judgment method for an automotive wire harness by a deep learning technique based on a convolutional neural network (CNN). In the proposed method, the CNN is used to find the relationship between the cross-sectional shape of the wire harness and a common mode current. We consider several patterns of the inputs and outputs of the CNN and discuss which input-output pattern is adequate for the judgment.