Weightless neural network based monitoring of screw fastenings in automated assembly

Abstract

Screw fastenings account for over a quarter of all assembly operations, and the intelligent automation of this process is of interest. This paper presents a new weightless neural network-based intelligent monitoring strategy for automated self-tapping screw insertions. A weightless neural network is designed and trained to monitor automated screw fastenings. The network is first trained and tested using computer simulations. The network is then tested on an experimental test setup, using both seen and unseen cases. Experimental results are presented to confirm the effectiveness of the approach. It is shown that the weightless neural network is relatively easy to train and is an efficient tool for monitoring automated screw fastenings.