Deep Neural Network-based Fingerprinting Localization for 5G NR mmWave Small-Cell

In this paper, we propose a fingerprinting-based localization method in the 5G millimeter-wave (mmWave) smallcell channel. The proposed method uses measurements that do not require additional processes to collect, such as synchronization signal-reference signal received power (SS-RSRP) used for synchronization between the user and base station in 5G communication and transmitter (TX) beam ID data as fingerprint data. With the collected data and a deep neural network (DNN)based pattern matching model, we evaluate the localization performance in 5G small-cell environments. As a result, the proposed method achieves localization root-mean-squared-error (RMSE) of 2.76m, which can be applicable to the actual smallcell environment.