Application of the artificial neural network method to detect defective assembling processes by using a wearable technology


KÜÇÜKOĞLU İ., ATICI ULUSU H., GÜNDÜZ T., Tokcalar O.

JOURNAL OF MANUFACTURING SYSTEMS, vol.49, pp.163-171, 2018 (Peer-Reviewed Journal) identifier identifier

  • Publication Type: Article / Article
  • Volume: 49
  • Publication Date: 2018
  • Doi Number: 10.1016/j.jmsy.2018.10.001
  • Journal Name: JOURNAL OF MANUFACTURING SYSTEMS
  • Journal Indexes: Science Citation Index Expanded, Scopus
  • Page Numbers: pp.163-171
  • Keywords: Industry 4.0, Wearable device, Artificial neural network, Signal classification, GLOVE, SENSOR

Abstract

Recently, the Industry 4.0 connects production processes and smart production technologies to lead up to a new technological age. The Industry 4.0 utilizes digital technologies such as augmented reality, sensors and wearables (e.g. smart watches, gloves, and glasses) to track all production operations. This study considers the problem of distinguishing proper and defective operations in connector assembly tasks in an automotive company. A digital assembly glove is developed as a wearable technology prototype. This glove is introduced to measure vibration and force values on the fingers to classify proper and defective operations in connector assembly processes. Experiments were conducted with 17 subjects to obtain force and vibration signals of the considered assembly task. For the signal classification of the digital assembly glove, the artificial neural network (ANN) methodology was used. Performance of the ANN was tested on the case of connector assembly process of the company. The collected proper and defective connection measurements were used for the training, validation, and testing of the ANN. As a result of the MATLAB computations, a neural network structure was obtained with 95% accuracy. The performance of the neural network showed that the ANN is an applicable method for the considered wearable technology to detect defective operations.