METHOD, COMPLEX FOR PROCESSING INFORMATION ABOUT FAILURES OF DEVICES OF WIRELESS SENSOR NETWORKS FOR DATA TRANSMISSION AND RELATED NETWORKS Russian patent published in 2023 - IPC H04L41/06 H04L41/16 

FIELD: communication systems.

SUBSTANCE: method and a device for handling failure messages from devices of a wireless sensor network for data transmission and related networks. The method comprises receiving, by means of a network element management system, the first alarm message with the first alarm association information containing the first wireless sensor network (WSN) object identifier, and receiving the second alarm message with the second association information containing the second identifier of the WSN object, and performing correlation analysis of the first alarm message and the second alarm message, while replacing the correlation analysis with the WSN object classification stage, and using the WSN object classifier, the WSN object membership relationship with the first and the second identifier to the class of sensors, or determine the relation of belonging of the WSN object with the first and second identifiers to the class of actuators, or determine the relation of belonging of the WSN object with the first and second identifiers to the component class, or determine the relation of belonging of the WSN object with the first and second identifiers to the sensor nodes class, or determine the relation of belonging of the WSN object with the first and second identifier to the gateways class, or determine the relation of belonging of the WSN object with the first and second identifier to the communication nodes class, or determine the relation of belonging of the WSN object with the first and second identifier to the backbone communication networks class, and use the stage of monitoring the technical condition of the classified objects of the WSN to determine the type of their technical condition, and determine the type of technical condition of the classified objects of the WSN with the first and second identifiers, and then use the stage of applying the rules of inference, which establish logical links between the classified objects of the WSN with the first identifiers and the type of their technical condition, and classified WSN objects with second identifiers and the type of their technical state, and get a logical conclusion about the initially failed device in the form of a classified WSN object with the first identifier and in an inoperative state, and simultaneously using the rules of inference, use stage of supervised machine learning with a classifier of classes of WSN objects, and use in the supervised machine learning model the membership relations of the WSN object with the first and second identifiers to the class of sensors, or to the class of actuators, or to the class of components, or to the class of sensor nodes, or to the class of gateways, or to the class of communication nodes, or to the class of backbone networks, and use information about the type of technical condition of the classified objects of the WSN, and obtain the output of supervised machine learning in the form of the first identifier of the classified object of the WSN, categorized as the primary source of messages about failures and being in an inoperable state, and comparing the output of machine learning with a teacher with the result of the actions of the inference rules, and determining the match of the first identifier of the WSN object obtained by the rules of inference and by the model of machine learning with a teacher, and determine the initially failed device.

EFFECT: improved reliability of identifying the initially failed device.

3 cl, 3 dwg