Proposal of a Smart Control System Using Edge Computing and Deep Learning Technologies

The core of this paper is a proposal of a control and monitoring system of the production process using smart sensors and Edge Computing (EC) technology. A solution proposed in this work provides the possibility of obtaining a wider spectrum of information than a normal control. Using the Multilayer perceptron neural network (MPL) and EC a faster data collection can be achieved without limiting the computing capacity and transfer services of the server. Additional benefits include availability of indirect data in process control, higher level of security by data collection and transfer, and, on top of that, low cost of components used. The proposed system reflects the current state of technology in the field, reflects on actual forms of the collection and usability of data, and offers opportunities to improve the quality of process management in general. This proposal shows that through EC, we are able to process a large amount of data (Big Data) from a number of devices/sensors (within proximity to the data), which we can analyze and process in a relatively short period of time. Thanks to analyzes and the use of AI (Deep Learning), we are able to react to errors or bad equipment conditions “Just in Time”.