Multi-agent technology of searching for digital radio-accelerated devices based on clustering by a bee colony method

Abstract

The article explores the capabilities of multi-position technology for searching of secret information retrieval devices based on bee colony clustering method. It has been found that the detection of such surveillance systems is becoming more and more difficult as their methods and modes of operation are also complicated. Possibilities of modern searching systems for the technical means of secret information retrieval are analyzed. It is shown that the existing means of detecting the hidden radiation over background of legal signals are ineffective and requires the introduction of intelligent methods of recognition. Considering the distributed nature of the problem, it is proposed to perform scanning through the involvement of several complexes, forming a physically multiagent environment in which the individual complexes will act as agents for the collection and processing of information. The bee colony method has been improved based on the use of a multiagent approach with direct agent-to-agent communication, which does not require a priori knowledge of the number of clusters and takes into account the specific features of secret information retrieval devices. An advanced clustering algorithm that can be implemented in a multiagent environment is proposed. The advanced method ensures that the best solution is matched by direct communication between agents, application of a natural selection procedure and an iterative procedure for searching the area by the agent. Experimental research was conducted by searching secret information retrieval devices in a room located in a multi-storey office building. A multiposition scanning complex was constructed from 3 Delta 4G complexes (each with 2 omnidirectional antennas) under common control. Field studies confirmed the increase of the clustering reliability by 6 ... 12% in comparison with the classical k-means method.