APPLICATION OF THE DEEP Q-NETWORK MACHINE LEARNING METHOD IN AN ADAPTIVE ANTENNA SYSTEM WITH AN ARTIFICIAL INTELLIGENCE UNIT

Abstract

The Deep Q-Network (DQN) machine learning method is analyzed. The possibility of applying the DQN machine learning method in an artificial intelligence module as part of an adaptive antenna system is shown in order to form a system of knowledge about the environment in which the adaptive antenna system operates. The algorithm for training a DQN agent is proposed and the proposed algorithm is implemented in the Python program code. As a result of the software implementation of the DQN algorithm, a knowledge system was obtained, which in the future will allow to realize the intelligent control of the radiation pattern petals of the adaptive antenna system.