ANALYSIS OF MACHINE LEARNING METHODS FOR SOLVING ANTENNA TECHNOLOGY PROBLEMS

Abstract

The scheme of modern machine learning methods classification is considered and presented. The content of each  methods in the classification is briefly disclosed. A structural diagram of an adaptive antenna system with an artificial intelligence unit is presented, and the principle of this scheme operation is explained. The role and function of the artificial intelligence unit within the adaptive antenna system are elucidated. The existing classes of machine learning methods are analyzed to determine their potential applicability to adaptive antenna systems for the purpose of intelligent control of the radiation pattern. It was determined that the primary machine learning method that could potentially be utilized by adaptive antenna system with artificial intelligence unit to achieve intelligent control of its directional characteristics is reinforcement learning. It is demonstrated that the reinforcement learning method can be integrated with other machine learning techniques, such as neural network methods, to enhance the precision of directing the directivity characteristics of adaptive antenna system with artificial intelligence unit.