Ground Level Mobile Signal Prediction Using Higher Altitude UAV Measurements and ANN

Abstract

Testing the mobile network signal strength is essential for evaluating actual user experience. This procedure is done by measurement campaign, where a person or a group of people walk or drive through the target area holding a measuring equipment. However, this is not suitable to do in hard-to-reach areas. In order to minimize human involvement and to reduce resources, labour, and time consumed, an alternative approach for physical assessment of cellular coverage and quality evaluating is needed. In this work, we used a drone to measure mobile network signal strength to generate a two-dimensional coverage map for difficult-to-reach areas. A machine learning algorithm is used to estimate the signal strength in other locations within the area to generate a dense 2D coverage map. The measurements were done on Sultan Qaboos University Campus, Muscat, Oman. Our finding shows that a drone equipped with a low-cost signal strength measuring device and an artificial neural network (ANN) algorithm are able to generate an accurate dense map of mobile signal strength in a flexible and cost-effective manner. The ANN was capable of predicting the signal strength at the ground from measurement at higher altitudes with an accuracy of 97%.