An Efficient Data-Driven Model for Millimeter-Wave 5G Channel Modeling Using Machine Learning and High-Performance Computing

Abstract

The fifth generation (5G) technology is efficiently designed to perform many things for the betterment of lives, such as Artificial Intelligence, Cyber-Physical Systems, the Internet of Things, etc. To facilitate this huge amount of data very high bandwidth is needed, hence 5G extensively uses the millimeter wave (mm-Wave) to enhance bandwidth. The technology of mm-Wave communication operates at very high frequencies, typically between 30 and 300 GHz. Some of the challenges will be key to realizing the full potential of mm-Wave communication technology for high-speed wireless communication in the future. The difficulties caused by mm-Wave are directivity, propagation loss, and sensitivity to blockage. To overcome these difficulties, we surveyed existing solutions and standards and identified research gaps. As a high data rate, mm-Wave may be considered in future generation communication and propagation channel requirements for mm-Wave investigated precisely for the prior knowledge of the quality of service (QoS) parameters. Therefore, channel modeling is the key need for the estimation of QoS parameters namely delay, angle of arrival, path loss, angle of departure, etc. In this paper, an efficient data-driven model for mm-Wave 5G Channel modeling using machine learning and high-performance computing is proposed which outperformed the other state-of-the-art in terms of various performance matrices.