INTELLIGENT ANALYSIS OF SENSORY DATA BASED ON FUZZY LOGIC AND NEURAL NETWORK
IN CRITICAL EVENT MONITORING SYSTEMS 

Anton A. Finogeev, Candidate of technical sciences, associate professor of the sub-department of computer-aided design systems, Penza State University (40 Krasnaya street, Penza, Russia), fanton3@yandex.ru

Background. The article deals with the issues of intellectual analysis of sensory data in a secure architecture of a system for monitoring critical events in the road infrastructure. Materials and methods. For intellectual analysis, clustering and packet recognition, a combined approach is proposed, which is implemented in the architecture of a software-defined network. The approach includes procedures for filtering invalid traffic, validating the authenticity of network nodes, a method of intellectual analysis of recognition and clustering of sensor data packets based on a fuzzy logic apparatus and a five-layer neural network in the space of nine features. The function of belonging to atypical packages is adjusted in the process of training the neural network, and the conclusion about the degree of similarity is formed using the fuzzy logic apparatus. Results. The architecture provides for a subsystem of intellectual analysis and clustering, which includes modules for validating network nodes and traffic filtering, a parser and packet analyzer based on fuzzy logic and a neural network,  and a logging module. Conclusions. The advantages of the proposed monitoring system are: software-defined network architecture, built-in network traffic analysis mechanisms, packet recognition and clustering method, network node validation method, invalid traffic filtering method from compromised nodes.

monitoring system, software defined network, sensor data, fuzzy logic, neural network, clustering

Finogeev A.A. Intelligent analysis of sensory data based on fuzzy logic and neural network in critical event monitoring systems. Modeli, sistemy, seti v ekonomike, tekhnike, prirode i obshchestve = Models, systems, networks in economics, technology, nature and society. 2023;(3):125–141. (In Russ.). doi: 10.21685/2227-8486-2023-3-9