Towards Intelligent Autonomous Controllers: Architecture for Industrial Distributed System

In industrial production systems, operators and engineers develop, on a daily basis, a degree of knowledge about some processes that, giving enough time, they become experts in the systems which themselves operate. In order to take advantage of this knowledge and experience, a knowledge-based system (KBS) can be employed to model and store the inference as well as the ways of decision-making of the operators in their working procedures. This article considers that in a setting of higher autonomy, a KBS could be embedded, such as for example, coupled in a programmable logic controller (PLC), which would adjust the controller as well as take control actions in a closed loop system without the direct intervention from the operators. Thus, as a possible application, such systems can be developed for the task of automatic tuning of proportional-integral-derivative (PID) parameters of a conventional controller, an important activity in process control operational routines. The study includes a conceptual proposal for developing this application and emphasizes the importance of artificial intelligence technologies for obtaining a higher degree of autonomy.