Comparison and assessment of machine learning approaches in manufacturing applications

Abstract

Abstract Machine learning (ML) is increasingly used in industry processes to advance digital technologies for Industry 4.0. This paper comprehensively reviews ML applications in manufacturing, covering supervised, unsupervised, and deep learning (DL) approaches across various industrial processes. The use of ML approaches in manufacturing process planning and control, fault identification/manufacturing/assembly, monitoring in the agricultural industry, quality control, and optimisation of logistics and robots are being investigated. Key highlights include an analysis of 70 primary studies, comparing recent trends in ML for manufacturing, and examining ML training concepts in learning factories. We also use ML techniques to assess the automotive manufacturing industry's architectures, models, and deployment challenges. Furthermore, these notions will be examined and applied to all possible approaches. The improvements in the scope of identification of the proper algorithm for the adequate set of applications will be examined further to ensure the smooth going of the process from training to the testing set.