Algorithm and Software for Intelligent Analysis of the Frequency Spectrum for Cognitive Radio Systems

Currently, systems with dynamic access to the radio spectrum technology, such as cognitive radio systems, are in demand and are actively developing. To ensure the effective operation of such systems, constant knowledge of the radio environment in the frequency range used is necessary. This article is devoted to the development of algorithms and software for intelligent analysis of the frequency spectrum to build a model for predicting the spectrum occupancy in time. The solution of this problem allows you to provide dynamic spectrum access without affecting the rights and quality of communication of primary users, while not reducing the quality of service for secondary users. This article has proposed and described an algorithm for implementing dynamic spectrum access with prediction based on spectrum occupancy models, consisting of 3 stages. Currently, software has been developed and a field experiment has been carried out to implement the first stage of the algorithm. The results of a field experiment for the first stage showed that the developed software allows us to identify free parts of the spectrum during the day. Also, the values of the duty cycle were obtained, allowing predicting the probability of occupancy of the frequency channel in the selected time period. Using this information, it is possible to build frequent (sequential) patterns to develop an up-to-date prediction model of spectrum occupancy.