IMPROVING THE METHOD OF SEARCHING DIGITAL ILLEGAL MEANS OBTAINING INFORMATION BASED ON CLUSTER ANALYSIS

Abstract

Іn the article the possibilities of the multipositional technology of searching digital insertion devices are investigated based on clustering. Existing means of detecting radiation of digital illegal means obtaining information reception show that they are ineffective on the background of legal signals of a multiagent medium. The constant improvement of digital illegal means obtaining information, masking their work under the signals of legal transmitters require the search for new approaches to the recognition and localization of these means. Prospects for the development of search technology today are associated with the creation of multi-position permanent detection and localization systems. However, the detection problem requires the recognition of harmful radiation on the set of statistics of signal parameters in the air by solving the problem of clustering. The disadvantages of most classical clustering methods are the need for prior knowledge of the possible number of clusters and a sufficiently high interactiveness, which complicates their practical application, especially in real-time. At the same time, intelligent multi-agent methods are free from these shortcomings, although their application remains quite complicated. The problem of recognizing the harmful signal against the background of similar legal signals is possible by using the method of a bee colony with direct communication between agents. In this case, the agents are individual elements of the multi-machine complex, which scan the ether at different points in space, then exchanging results with other agents, and, finally, come to a common conclusion about the nature of the signal. Full-scale studies have been carried out that confirm the reliability of clusterization by 6 ... 12% compared to the classic k-medium method.

Keywords

• information protection, inbound device, multiagent system, clusterization