An intelligent framework for modelling and simulation of artificial neural networks (ANNs) based on augmented reality
- PDF / 3,985,731 Bytes
- 14 Pages / 595.276 x 790.866 pts Page_size
- 20 Downloads / 207 Views
ORIGINAL ARTICLE
An intelligent framework for modelling and simulation of artificial neural networks (ANNs) based on augmented reality D. Mourtzis 1
&
J. Angelopoulos 1
Received: 12 August 2020 / Accepted: 28 September 2020 / Published online: 15 October 2020 # Springer-Verlag London Ltd., part of Springer Nature 2020
Abstract The digitalization of industry is targeting at the integration of artificial intelligence (AI) in manufacturing systems, for delivering intelligent machinery. Although AI seems a long-term target, similar enabling technologies such as artificial neural networks (ANNs) have been introduced. Despite that ANNs are inspired by the human brain’s functioning, understanding how they work and training them is a challenging task, requiring engineers with advanced math and coding skills. On the contrary, augmented reality (AR) is a cutting-edge digital technology, enabling the registration of 3D content on the physical environment, thus enhancing user’s perception in a growing variety of scientific fields. Therefore, this research work aims at the design and development of an AR-based framework that facilitates the conceptualization of an ANN through AR, assists engineers train efficient ANN and moreover share knowledge through suitable communication channels. Finally, the framework can handle datasets with the use of cloud services. Keywords Augmented reality . Artificial neural network . Modelling . Simulation . Cloud services
1 Introduction In the era of vast digitalization, big volumes of information are constantly being produced. According to the survey presented in [1], about 80% of the data produced within the premises of a company are unstructured. Extending the previous statement, the study presented in [2] reflects the quantity of data produced in a manufacturing plant. It is undeniable that such volumes of data cannot be processed by the available human resources. In addition to that, even if humans were capable of doing such task, it would require their full commitment on this task, for the rest of their “working life”. Hence, engineers around the world have focused their research activities on new technological topics of data science. One of these topics is about artificial neural networks (ANNs). ANN is the enabling technology of mimicking human brain functionalities supported by advanced computational power, which is readily
* D. Mourtzis [email protected] 1
Laboratory for Manufacturing Systems and Automation, Department of Mechanical Engineering and Aeronautics, University of Patras, Rio, 26504 Patras, Greece
available, thanks to the ongoing technological advances under the umbrella of Industry 4.0. To that end, ANN is thought to be the new trend of modern manufacturing systems. According to a relevant report [3], machine learning adoption has risen from 58% in 2017 to 63% in 2018. Similarly, deep learning adoption marked a 16% raise for the same time span, and natural language processing adoption reached 62% in contrast to the 53% for 2017. Finally, computer vision ca
Data Loading...
