Polygon for smart machine application

The article deals with the implementation of the paradigm “a dependable and safe system out of undependable and not entirely safe components” using the example of application of complex situation monitoring, where global situation model is required. The description of the situation is based on sensory data from various locations. When it is difficult to identify the situation on the basis of incomplete and unreliable sensor systems, additional information is obtained about the state of these locations using a small group of mobile smart machines. Increasing the reliability of the situation description is achieved by generalizing, abstracting and reasoning based on heterogeneous sensory data, but not because of their duplication. The article deals with the training and research polygon of intelligent information technologies for approbation of such applications of intelligent machines. The architecture of the polygon, components of sensory systems, organized in the form of IoT locations, and smart machines are presented. It deals with the organization of the software supporting, also. A description of two projects implemented at the polygon is provided: monitoring of fire-dangerous situations and monitoring of dangerous situations related to the violation of order in the campus auditoriums. In both applications, smart machines are implemented on the basis of training wheeled robots, equipped with heterogeneous sensor systems. Software and hardware implementation is performed on Raspberry Pi, Arduino and ESP8266 platforms.