Extraction of knowledge from multimodal flows of unstructured data on the basis of self-organization of multi-agent cognitive architecture of the mobile robot

The problem of knowledge acquisition from unstructured multi-modal data streams is considered in the context of the construction by the mobile robot, immersed in a real environment, the epistemology, allowing it to achieve the successful implementation of the desired functionality. The recursive cognitive multi-agent architecture is used as the basis of adaptive robot control system. Self-organization of that architecture results in the construction of causal dependencies, expressed in terms of production rules, forming in the knowledge bases of agents. A program for 3D-visualization has been developed. An example of the construction of the epistemology of the robot based on multi-modal data is presented.