INTELLIGENT SENSOR NETWORKS IN INDUSTRY 5.0. A GENERALIZED CONCEPT OF CREATING DIGITAL PLATFORMS FOR MANAGING COMPLEX SYSTEMS BASED ON A REGULARIZING BAYESIAN APPROACH (report at the international conference on soft computing and measurements SCM-2022)

UDC 621
DOI: 10.36871/2618-9976.2022.05.003

Authors

Prokopchina S.V.
Doctor of Technical Sciences, Professor, Financial University under the Government of the Russian Federation, Moscow, Russia

Abstract

The concept of Industry XNUMX provides for the intensive development of the processes of intellectualization of sensor systems. Among the most important specific properties of real measuring processes in complex systems is, first of all, their implementation under conditions of significant uncertainty. Uncertainty is due to a priori incompleteness, inaccuracy, vagueness of information about a complex measuring object and its operating environment, which does not allow building an adequate model of the object before conducting a measuring experiment, identifying and formalizing the influencing environmental factors and developing effective algorithms for the functioning of information and measuring systems. The article proposes an approach to the intellectualization of measuring systems under conditions of uncertainty by creating intelligent sensor networks based on Bayesian intelligent technologies (BIT) and means of their implementation. Typical modules of such networks are considered, which are integrated sets of various sensors and intelligent systems for processing measurement information. Such sensor sets can include both physically implemented measuring devices and virtual sensors for measuring nonquantitative or integral characteristics. The results of the work of networks are complex assessments of the state of complex objects and recommendations for ensuring their sustainable functioning. An important part of such systems is the builtin means of complete metrological substantiation of all obtained solutions. The systems have a hierarchical architecture corresponding to the levels of control of complex objects, which has the possibility of selfdevelopment based on newly received information. This is achieved thanks to models and measuring scales with dynamic constraints, on which all the algorithms used in these networks are built. The article provides examples of the use of intelligent sensor networks for monitoring and managing technical and socioeconomic systems.

Keywords

Smart measurements, Bayesian approach, Sensor networks