Machine Activity Recognition Using Clustering Method

Abstract

Machine activity recognition is important for benchmarking and analysing the performance of individual machine, machine maintenance needs and automated monitoring of work progress. Additionally, it can be the basis for optimizing manufacturing processes. This article presents an attempt to use object clustering algorithms for recognizing the type of activity in the production complex. For this purpose, data from the production process and the k-means algorithm were used. The most common object clustering algorithms were also discussed. The results and the presented analysis approach demonstrate that this method can be successfully utilized in practice.