Machine learning based screw drive state detection for unfastening screw connections

Anwar Al Assadi, David Holtz, Frank Nägele, Christof Nitsche, Werner Kraus, Marco F. Huber

Abstract: Highlights•The strength of this approach is the ML-based detection of stripped screws (Independent of screw head drive and screw size).•Multi-layer perceptrons (MLPs) and convolutional neural networks (CNNs) are suitable for screw drive state detection.•The findings of this paper improve the state-of-the-art screw drive detection in terms of required revolution numbers. (720°−> 360°)•ML techniques are very promising for decision-making process in automated unfastening screw connections.

External IDs:doi:10.1016/j.jmsy.2022.07.013