Large Scale Channel Parameters Estimation by Utilization of Reflected Rays Information for an Urban Environment

In recent years, the discovery and utilization of the millimeter-wave for the next generation of mobile communications have fuelled the demand to develop site specific and more accurate wireless channel models. Due to the dynamic and high interactive nature of the millimeter waves, accurate models which takes into consideration the effects of the interacting objects in the propagation environments are needed. This paper presents a development of a machine learning (ML) based prediction model for large scale channel parameter (LSCP) estimation by analyzing the interaction of the reflected multipath rays information with propagation environments. This is to capture the shape of the changing propagation environments more accurately. Utilizing the robustness of ML, a dynamic LSCP prediction model can be realized. The input parameters to the prediction model are transmitter (TX) and receiver (RX) positional coordinates, and the reflected rays’ information such as the delay, angle of arrival, angle of departure, elevation angle of arrival, elevation angle of departure, and power gain. The proposed model was implemented using Random Forests. Ray tracing (RT) simulation was performed to calculate the input dataset of the LSCPs, and also the input datasets of the reflected rays. Cross validation was utilized to validate the model.