Deep Neural Network for Localization of Mobile Users using Raytracing

As the world evolves towards faster data transmission, there is an ever-increasing demand for better user localization which will find application in transport, medicine, and robotics. In this study, we present an accurate localization algorithm for mobile users using deep neural network with Bayesian optimization and a communication channel operating at 3.75GHz frequency. We design a deep neural network model which facilitates and speeds up the localization process. The deep neural network (DNN) is utilized in this study to locate moving user equipment (UEs) with randomly assigned velocities. Using preliminary computer simulations, we present a method for training a neural network that extracts channel parameters (features) that are used to estimate location. Our method produces localization accuracy for line of sight (LOS) users, less than 1 m error, and the accuracy can further be improved by implementing higher rate of data sets.