Intelligent Base Station Placement in Urban Areas With Machine Learning

Abstract

This paper presents a time-reliable machine learning model for accurate and rapid base station placement in urban areas. Several real-world city data are used for training the machine learning model to predict parameters such as path loss values. It is developed using supervised training and multiple regressions with the use of artificial intelligence. Parameters such as coverage dimensioning, bandwidth of frequencies, building's density and height, propagation model tuning and path loss values are used to predict the locations of base stations. Once the predicted base station position is obtained, the average percentage error is calculated against the real-world data of base station that shows an accuracy of 84%. This proves that our model is a reliable tool to predict future base station locations. This model will benefit the telecommunication industries to reduce time and cost for rapid base station planning in urban areas that see frequent changes to their landscapes.