Machine-Learning-Based Ground-Level Mobile Network Coverage Prediction Using UAV Measurements

Future mobile network operators and telecommunications authorities aim to provide reliable coverage. Signal strength, normally assessed using standard drive tests over targeted areas, is an important factor strongly linked user satisfaction. Drive are, however, time-consuming, expensive, can be dangerous in hard-to-reach areas. An alternative safe method involves drones or unmanned aerial vehicles (UAVs). The objective of this study was use a drone measure signal strength at discrete points few meters above the ground artificial neural (ANN) for processing measured data predicting level. equipped with low-cost logging equipment. ANN also used classify specific locations terms coverage into poor, fair, good, excellent. training testing were collected by measurement unit attached different areas Sultan Qaboos University campus Muscat, Oman. A total 12 topologies scanned. proposed achieved accuracy 97% level based on measurements taken higher altitudes. In addition, performance evaluated several test scenarios, achieving less than 3% mean square error (MSE). Additionally, angles respect vertical tested, prediction MSE found approximately angle 68 degrees. outdoor predict indoor 6%. Furthermore, attempt find globally accurate module area, all zones’ cross-tested modules trained zones. It that, within tested 10% from only one zone.