Federated learning based base station selection using LiDAR data

Abstract

The optimum selection of base station (BS) around the mobile vehicle is important for establishing communication links in the mm-Wave-based communication systems to get a reliable and low latency link between them. For selecting the best BS, each BS performs a handshake with the mobile vehicle, using the ray tracing method power loss is calculated between them and the best BS is selected. In this research, we have investigated the best base station out of the 3 BSs we have considered in our system model using Ray tracing method and RSSI values and further compared it with that calculated using machine learning (ML) model. We have also applied federated learning (FL) algorithms to our ML model to reduce the communication overhead and to preserve the privacy of the user. Different FL algorithms are compared based on various parameters in our model to get the test accuracy of these algorithms, where the simulation results shows the accuracy achieved using these algorithms.