Mathematical model of optimal reliability management in security systems with neural network controllers

Abstract

Mathematical model of optimal reliability management in security systems with neural network controllers

Tsarkova E.G.

Incoming article date: 05.03.2022

The paper discusses methods for improving the efficiency of perimeter security systems through the use of neural network signal processing technologies that serve to increase the information content of data coming from security detectors. The proposed approach allows not only to establish in a timely manner the fact of illegal actions to overcome the perimeter of the object, but also to classify the violator with a minimum level of false positives. Special attention is paid to the issues of optimal control of the reliability of signal processing units with a logical neural network analyzer, considered as queuing systems.

Keywords: integrated security, security detector, detection zone, signal processing unit, artificial neural network, neural network signal processing, reliability, queuing system, optimal control, maximum principle