A CNN Technique for MCS Selection in 5G NR Mobile Communication Systems.

5G NR is a wireless mobile communication system that supports ultra-high speed mobile communications, high reliability services, and large-scale Internet of Things. In a mobile communication environment, the Doppler effect, which is proportional to the speed of travel, causes channel changes over time, which can lead to communication performance degradation. In order to achieve optimal communication performance in a mobile communication environment, it is necessary to predict the Signal to Noise Ratio (SNR) between the base station and the terminal and select and transmit the most appropriate Modulation and Coding Scheme (MCS) accordingly. In this paper, we propose a method for selecting the MCS level of a single antenna based on Convolutional Neural Network (CNN) in 5G NR mobile communication systems. The proposed system assumes a time division duplex (TDD) scheme, measures the SNR at the time of reception, predicts the SNR at the time of future transmission using CNN based on the measured past channel information, and selects the MCS level based on the predicted SNR. Experimental results through computer simulation show that the proposed CNN-based MCS selection method has a lower probability of communication disconnection and higher transmission rate at all speeds compared to the existing average value method based on the average of SNR and the recent value method based on the most recently received SNR. In particular, the transmission speed of the proposed method is about 46% and 4.6% better than the existing average value method and recent value method, respectively, and can be utilized as a technology to increase the transmission speed in 5G mobile communication environment. [ABSTRACT FROM AUTHOR]