Systematic Literature Review on the Machine Learning Techniques for UAV-Assisted mm-Wave Communications

The application of machine learning (ML) solutions in UAV-assisted 5G communication can bring significant benefits to 5G and beyond 5G communication. There is little elementary, secondary, and tertiary study on machine learning applications in UAV-assisted 5G communication. The apparent paucity of such investigation makes it hard to develop precise solutions for UAV-assisted 5G communication. Therefore, it is essential to study and comprehend how to use machine learning in UAV-aided 5G communication. In this article, we deliver a systematic study of all-important research activities, in which machine learning (ML) methods have been used on the wireless communication based on UAV for improving several design and functional characteristics such as beamforming, resource allocation, dynamic deployment, and trajectory prediction. The studies were clustered into four themes: the main machine learning algorithms applied in UAV-assisted wireless communication; UAV-assisted 5G communication process in which machine learning processes and/or frameworks are applied; main application categories of machine learning algorithms and/or frameworks; and the results of this review specify that the main machine learning algorithms/framework applied in UAV-assisted wireless communication are: Q-Learning, MARL, K-means, AMSSA, genetic algorithm, support vector machine (SVM), support vector regression, artificial neural network (ANN), LSM, cross-entropy algorithm, DL algorithm, and reinforcement learning algorithm.