Biological Principles of Integration Information at Big Data Processing

The article is devoted to the principles of integration of heterogeneous information in order to obtain generalized assessments of the state of various objects and systems. A characteristic feature of cyber-physical systems is the generation and availability of large volumes of heterogeneous information. The value of this information is entirely determined by its processing mechanisms. So for monitoring problems, the depth of processing and the semantic context of the processed data are crucial for diagnosing and predicting the state. The mechanisms of information integration are particularly important when building a digital twin, which is a virtual image of an object or system. In the paper, from the standpoint of the Bernstein's theory of movement construction levels, an analysis of the mechanisms of information integration in the formation of internal behavior patterns is carried out. The modern hypotheses of the formation of polysensory models and the mechanisms of their invariant representations are separately considered. The principles of the integration of heterogeneous information in the big data processing are formulated. The possibility of implementing these principles using the apparatus of cellular automata is shown.